temporal changes affect plant chemistry and tritrophic interactions

13
Basic and Applied Ecology 8 (2007) 421—433 Temporal changes affect plant chemistry and tritrophic interactions R. Gols a, , C.E. Raaijmakers b , N.M. van Dam b , M. Dicke a , T. Bukovinszky a , J.A. Harvey b a Laboratory of Entomology, Department of Plant Sciences, Wageningen University, P.O. Box 8031, 6700 EH Wageningen, The Netherlands b Department of Multitrophic Interactions, Netherlands Institute of Ecology, P.O. Box 40, 6666 ZG Heteren, The Netherlands Received 13 March 2006; accepted 11 September 2006 KEYWORDS Brassica oleracea; Diadegma semiclausum; Direct defence; Glucosinolates; Host suitability; Plutella xylostella; Sinapis alba Summary In nature, individuals of short-lived plant species (e.g. annuals, biennials) may grow at different times during the growing season. These plants are therefore exposed to different season-related conditions such as light and temperature, which in turn may have consequences for primary and secondary chemistry of the plant. Despite this, many studies examining plantconsumer interactions are performed in single replicates, which may thus ignore temporal variation in the expression of phenotypic plant traits that affect multitrophic interactions. In the present study, we demonstrated that even under strictly controlled conditions in a greenhouse, secondary plant chemistry changes dramatically in plants growing at different times in a single year, i.e. July, August and November. Glucosinolate (GS) contents in herbivore-damaged leaves of two different crucifer species, Brassica oleracea and Sinapis alba were higher in the August and November replicates than in the July replicate and GS concentrations were 1025 times higher in S. alba than in B. oleracea. The development of a specialist herbivore, Plutella xylostella, also varied significantly over the three replicates. Larvae of P. xylostella that had fed upon either S. alba or B. oleracea, attained the largest biomass and had the fastest development rate in the November replicate. Female P. xylostella moths grew larger on S. alba than on B. oleracea, whereas male biomass was not significantly affected by host-plant species. Plant species, but not season also affected performance of the endoparasitoid, Diadegma semiclausum. Similar to the performance of host females, parasitoid males developed faster and attained the highest biomass when attacking P. xylostella larvae feeding on S. alba. Most importantly, the performance of the herbivore and its parasitoid only appeared to partially conform to levels of GS in ARTICLE IN PRESS www.elsevier.de/baae 1439-1791/$ - see front matter & 2006 Gesellschaft fu ¨r O ¨ kologie. Published by Elsevier GmbH. All rights reserved. doi:10.1016/j.baae.2006.09.005 Corresponding author. Tel.: +31317 482330; fax: +31317 484821. E-mail address: [email protected] (R. Gols).

Upload: r-gols

Post on 04-Sep-2016

213 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: Temporal changes affect plant chemistry and tritrophic interactions

ARTICLE IN PRESS

Basic and Applied Ecology 8 (2007) 421—433

1439-1791/$ - sdoi:10.1016/j.

�CorrespondE-mail addr

www.elsevier.de/baae

Temporal changes affect plant chemistry andtritrophic interactions

R. Golsa,�, C.E. Raaijmakersb, N.M. van Damb, M. Dickea,T. Bukovinszkya, J.A. Harveyb

aLaboratory of Entomology, Department of Plant Sciences, Wageningen University, P.O. Box 8031,6700 EH Wageningen, The NetherlandsbDepartment of Multitrophic Interactions, Netherlands Institute of Ecology, P.O. Box 40, 6666 ZG Heteren,The Netherlands

Received 13 March 2006; accepted 11 September 2006

KEYWORDSBrassica oleracea;Diadegmasemiclausum;Direct defence;Glucosinolates;Host suitability;Plutella xylostella;Sinapis alba

ee front matter & 2006baae.2006.09.005

ing author. Tel.: +31 317ess: [email protected] (

SummaryIn nature, individuals of short-lived plant species (e.g. annuals, biennials) may growat different times during the growing season. These plants are therefore exposed todifferent season-related conditions such as light and temperature, which in turn mayhave consequences for primary and secondary chemistry of the plant. Despite this,many studies examining plant–consumer interactions are performed in singlereplicates, which may thus ignore temporal variation in the expression of phenotypicplant traits that affect multitrophic interactions. In the present study, wedemonstrated that even under strictly controlled conditions in a greenhouse,secondary plant chemistry changes dramatically in plants growing at different timesin a single year, i.e. July, August and November. Glucosinolate (GS) contents inherbivore-damaged leaves of two different crucifer species, Brassica oleracea andSinapis alba were higher in the August and November replicates than in the Julyreplicate and GS concentrations were 10–25 times higher in S. alba than inB. oleracea. The development of a specialist herbivore, Plutella xylostella, alsovaried significantly over the three replicates. Larvae of P. xylostella that had fedupon either S. alba or B. oleracea, attained the largest biomass and had the fastestdevelopment rate in the November replicate. Female P. xylostella moths grew largeron S. alba than on B. oleracea, whereas male biomass was not significantly affectedby host-plant species. Plant species, but not season also affected performance of theendoparasitoid, Diadegma semiclausum. Similar to the performance of host females,parasitoid males developed faster and attained the highest biomass when attackingP. xylostella larvae feeding on S. alba. Most importantly, the performance of theherbivore and its parasitoid only appeared to partially conform to levels of GS in

Gesellschaft fur Okologie. Published by Elsevier GmbH. All rights reserved.

482330; fax: +31 317 484821.R. Gols).

Page 2: Temporal changes affect plant chemistry and tritrophic interactions

ARTICLE IN PRESS

R. Gols et al.422

damaged leaves, indicating that there is a complex of factors involved indetermining the effects of plant quality on higher trophic levels.& 2006 Gesellschaft fur Okologie. Published by Elsevier GmbH. All rights reserved.

ZusammenfassungIn der Natur konnen Individuen von kurzlebigen Pflanzenarten (z.B. Annuelle,Bianuelle) zu verschiedenen Zeiten in der Saison wachsen. Diese Pflanzen sind daherunterschiedlichen jahreszeitlichen Bedingungen wie Licht und Temperatur ausge-setzt, was wiederum Konsequenzen fur die primare und sekundare Chemie derPflanze haben kann. Trotzdem sind viele Untersuchungen, welche die Pflanzen-Konsumenten-Interaktionen untersuchen, einfache Ansatze, die so die zeitlicheVariation in der Expression von Pflanzenmerkmalen ignorieren, die Einfluss aufmultitrophische Interaktionen hat. In der vorliegenden Untersuchung zeigten wir,dass sich selbst unter den strikt kontrollierten Bedingungen im Gewachshaus diesekundare Pflanzenchemie bei Pflanzen, die zu verschiedenen Zeiten in einemeinzigen Jahr, d.h. im Juli, August und November, wuchsen, dramatisch andert. DieGehalte an Glukosinolat (GS) in den von Herbivoren beschadigten Blattern von zweiKreuzblutlerarten, Brassica oleracea und Sinapis alba, waren im August- undNovember-Ansatz hoher als im Juli-Ansatz und die GS-Konzentrationen waren beiS. alba zehn bis zwanzigfach hoher als bei B. oleracea. Die Entwicklung desspezialisierten Herbivoren Plutella xylostella, variierte ebenfalls signifikant bei dendrei Ansatzen. Die Larven von P. xylostella, die sowohl auf S. alba als auch aufB. oleracea fraßen, erreichten beim November-Ansatz die großte Biomasse undhatten die schnellste Entwicklungsrate. Weibliche P. xylostella Falter wuchsen aufS. alba schneller als auf B. oleracea, wahrend sich die Biomasse der Mannchen nichtsignifikant zwischen den Wirtsarten unterschied. Die Performanz des EndoparasitenDiadegma semiclausum wurde von der Pflanzenart, nicht aber von der Jahreszeit,beeinflusst. Vergleichbar zur Performanz der Wirtsweibchen entwickelten sich dieParasitoidenmannchen schneller und erreichten die hochste Biomasse, wenn sieP. xylostella Larven befielen, die auf S. alba fraßen. Vor allem aber schien diePerformanz des Wirtes und seiner Parasitoiden nur teilweise mit den Gehalten vonGS in den beschadigten Blattern ubereinzustimmen, was darauf hinweist, dass eseinen Komplex von Faktoren gibt, die den Einfluss der Pflanzenqualitat auf hoheretrophische Ebenen bestimmen.& 2006 Gesellschaft fur Okologie. Published by Elsevier GmbH. All rights reserved.

Introduction

Plants have evolved a range of physical (tri-chomes, spines, wax layers) and chemical (toxins,repellents and digestibility reducers) barriers toprevent or reduce insect feeding (reviewed byKarban & Myers, 1989; Schoonhoven, van Loon, &Dicke, 2005), which are collectively known asdirect plant defenses. Many insects have evolvedcounter-adaptations to circumvent these plantdefenses. They can avoid feeding on toxic plantparts or tissues (Dussourd, 1993), and some insectscan excrete or detoxify harmful plant compounds(e.g. Naumann, Hartmann, & Ober, 2002; Ratzka,Vogel, Kliebenstein, Mitchell-Olds, & Kroymann,2002). In addition, herbivores can sequester planttoxins and use them for their own defence (e.g.Muller et al., 2001). Many herbivores have evolveddietary specialization and feed only on one or a fewrelated plant species. The performance of specia-

lized herbivores is often determined by thepresence of specific chemicals in the host plantthat stimulate feeding balanced against the nega-tive effect of inhibitors (Gupta & Thorsteinson,1960a; Reed, Pivnick, & Underhill, 1989; Renwick,2002; van Loon & Schoonhoven, 1999). Somesecondary plant chemicals in potential host plantscan negatively affect the development of specia-lized herbivorous insects (Agrawal & Kurashige,2003; Steppuhn, Gase, Krock, Halitschke, & Bald-win, 2004; Stowe, 1998; van Dam, Hadwich, &Baldwin, 2000; Zangerl & Berenbaum, 1993). Plant-defense chemicals are also transferred verticallythrough the food chain and can affect the perfor-mance of organisms in higher trophic levels(Barbosa, Gross, & Kemper, 1991; Campbell &Duffey, 1981; Francis, Lognay, Wathelet, & Hau-bruge, 2001; Hartmann, 2004; Harvey, van Dam, &Gols, 2003; Roth, Knorr, & Lindroth, 1997; Sznajder& Harvey, 2003).

Page 3: Temporal changes affect plant chemistry and tritrophic interactions

ARTICLE IN PRESS

Temporal changes affect tritophic interactions 423

Plant secondary chemistry is genetically deter-mined, but is phenotypically plastic and varies inresponse to both biotic (e.g. pathogen infectionand herbivore feeding, see Agrawal, 1999; Karban& Baldwin, 1997) and abiotic (e.g. nutrient levelsand light conditions, Gouinguene & Turlings, 2002;Stout, Brovont, & Duffey, 1998; Takabayashi, Dicke,& Posthumus, 1994; Traw & Dawson, 2002; van Damet al., 2000) factors. Seasonal effects on secondarychemistry of ephemeral plants, such as annuals,have received little attention in the literature(Feeny & Rosenberry, 1982). However, many short-lived plants may germinate and grow at differenttimes during the growing season; for example,Sinapis arvensis can be found between March andDecember in temperate latitudes (J.A. Harvey,personal observation). Abiotic factors, such as lightand temperature, and the intensity of selectionpressures from attackers, such as pathogens andherbivores, may also change dramatically duringthe year. Consequently, these changes may havesignificant effects on the primary and secondarychemistry of cohorts of annual plants that germi-nate at different times during the year. Differentgenerations and/or guilds of herbivores associatedwith these plants may also be affected over theseasons in which these plants grow. However,laboratory studies, carried out under controlledconditions, are often performed in single replicatesand at any time of the year. They assume that theplant will exhibit standard phenotypic responsesunder these controlled conditions. Thus, whereasplant age is usually well defined in experimentalstudies, any possible effects of season in green-house-reared plants are usually ignored.

All plant species belonging to the Brassicaceaefamily are able to biosynthesize glucosinolates (GS)(b-thioglucocide-N-hydroxysulfates) (Fahey, Zalc-mann, & Talalay, 2002; Mithen, 2001). GS areconverted into biologically active thiocyanates,isothiocyanates, nitriles and oxazolidine-2-thionesby the enzyme myrosinase, which is stored andsegregated from the GS substrate and is onlyreleased after wounding of the plant. GS and theirbreakdown products are considered to play animportant defensive role against pathogens andinsect herbivores. Members of the Brassicaceaevary considerably in GS quantity and composition(Fahey et al., 2002; Rask et al., 2000) and insectherbivores specialized on Brassicaceae differ intheir performance on different members of thisplant family (Fox, Kester, & Eisenbach, 1996;Ohsaki & Sato, 1999; Sznajder & Harvey, 2003).Differences in GS quantity and composition be-tween crucifer species may thus contribute todifferences in herbivore performance.

Plutella xylostella L. (Lepidoptera, Plutellidae)is a specialist herbivore that is known to feed on anumber of species in the Brassicaceae. Adultfemales and larval stages use GS as ovipositionand feeding stimulants, respectively (Gupta &Thorsteinson, 1960a, b; Nayar & Thorsteinson,1963; Pivnick, Jarvis, & Slater, 1994). However, anincrease in GS concentration does not alwayspositively correlate with larval performance(Li, Eigenbrode, Stringham, & Thiagarajah, 2000).

The aim of the present study is to evaluate host-plant quality over the course of a growing seasonfor the herbivore P. xylostella feeding on twocrucifer species that differ qualitatively and quan-titatively in GS content, and to examine if there arealso effects on the development of its solitaryendoparasitoid, Diadegma semiclausum Hellen (Hy-menoptera: Ichneumonidae). The experimentswere conducted over several months in a green-house with a controlled temperature regime butwith seasonal (temporal) changes in light accessi-bility. Under these conditions, we predicted thatplant quality, based on levels of allelochemicals inplant tissues, would vary only slightly over theduration of the experiment. Because of this, weexpected that the effects of marginal differences inplant defences on the herbivore and its endopar-asitoid would be negligible but that insect perfor-mance would still be correlated with the levelsof GSs.

Our results reveal that even under fairly con-trolled temperature conditions, there are dramatictemporal (seasonal) shifts in foliar levels of plantallelochemicals. However, herbivore and parasitoidperformance was not strictly correlated with plantdefences revealing that other factors are involvedthat affect their development. We argue that thevariation and expression of abiotic and biotcstresses in the field are likely to be far larger thanin the laboratory, and that there may be significantconsequences for the biology and ecology ofgeneralist and specialist consumers over severaltrophic levels.

Materials and methods

Insects

Larval and adult P. xylostella were obtained fromcultures maintained in the laboratory on Brassicaoleracea plants (cv. Cyrus). The parasitoid,D. semiclausum was collected from Brussels sproutfields in the vicinity of Wageningen (The Nether-lands) and was reared on larvae of the diamond-back moth, P. xylostella. A new generation of

Page 4: Temporal changes affect plant chemistry and tritrophic interactions

ARTICLE IN PRESS

R. Gols et al.424

D. semiclausum was started weekly. About 40 waspsin a 50:50 male-to-female sex ratio were added to acage with two B. oleracea plants heavily infestedwith second and third instar P. xylostella larvae.Extra food plants were added regularly for thecaterpillars, as well as honey and water forthe parasitoids. Two weeks after introduction ofthe wasps, all plant material was removed andcocoons were collected and transferred to a cleaninsect cage (37� 40� 30 cm) without plants. Emer-ging wasps were collected daily and used either forparasitoid rearing or experiments. The cultures ofP. xylostella and D. semiclausum were maintainedin a climate room, 16L:8D photoperiod at 2172 1Cand a relative humidity of 70%. Wasps werecollected over a 3-week period from three differentgenerations and kept collectively in a cage(42� 22� 22 cm) at 1272 1C to extend their long-evity. Wasps were provided ad libitum with waterand honey, which was administered to the gauzewall of the cage.

Plants

In cultivated B. oleracea L. plants, the mostdominant GS is the indole GS glucobrassicin,whereas in Sinapis alba L. the aromatic sinalbin isthe major GS component (Francis et al., 2001;Hopkins, Ekbom, & Henkow, 1998; Reed et al.,1989; Renwick & Lopez, 1999).

B. oleracea L. gemmifera cv. Cyrus (Brusselssprout) and S. alba L. cv. Carneval (white mustard)were grown in a greenhouse (2575 1C, 40–80%relative humidity and a photoperiod of at least16 h). If the light dropped below 500 mmol photons/m2/s during the 16-h photoperiod, supplementaryillumination was applied by high-pressure mercurylamps. Seeds were sown individually in pots (11�11� 11 cm) filled with potting soil (‘‘Lentse pot-grond’’ no. 4, Lent, The Netherlands). When plantswere used for oviposition by P. xylostella, mustardplants were 3–4 weeks old and Brussels sproutplants were 5–6 weeks old. Brussels sprout plantswere sown on June 4, July 7 and October 1, andsowing dates for mustard were June 18, July 18 andOctober 13, 2002.

Bioassay

We monitored the development of parasitizedand unparasitized P. xylostella larvae that werefeeding on either B. oleracea or S. alba. Toinvestigate the effect of instar at parasitism onwasp development, host larvae were parasitized byD. semiclausum in either the second (L2) or fourth

(L4) larval instar (see below). To obtain larvalstages of P. xylostella on the two different plantspecies, two plants of each species were placed inseparate cages per plant species with c. 100P. xylostella moths in a 50:50 female-to-male ratio.After a 24 h oviposition period, the moths wereremoved and the plants were left in the cages forthe development of the eggs into larvae. Extraplants were added to the cages to sustain larvalgrowth. When the larvae reached the second instar,a cohort of 70–100 unparasitized caterpillars wastransferred to new plants and a second cohort of100–110 individuals was parasitized by D. semi-clausum. An equal number of P. xylostella larvaewas parasitized when the unparasitized larvaereached the fourth instar.

For parasitism, a vial with a single mated femalewasp was presented with a leaf with a single hostlarva. After the larva was seen to be parasitized, itwas transferred to its respective experimental hostplant with a fine paintbrush. Each female wasp wasused to parasitize up to six individual hosts.Parasitized larvae were evenly distributed overundamaged host plants, 10 plants for mustard andnine plants for Brussels sprout (Brussels sprout hadabout 10% more biomass than mustard plants). Eachindividual plant received 10–12 larvae. Plants of thesame species and the same treatment (parasitized asL2, as L4, or unparasitized) were placed together in acage, (60� 60� 100 cm). The larvae were allowed tomove freely from one plant to another within theirrespective cage. When the larvae reached the fourthinstar, pieces of corrugated cardboard (5� 20 cm)were placed on top of the plants as the larvae tendto pupate in sheltered places. This prevented thelarvae pupating directly on the cage walls. Pupaewere carefully cut out from the cardboard orremoved from the leaves on which they had pupatedand transferred individually to labelled glass vials.When adults emerged, either as moths or wasps, theegg-to-adult development time was recorded. Vialswere checked every 2h, but development time wasrecorded in full days, since the exact time ofoviposition had not been recorded. Adults werefrozen and dried for 72h in an oven at 80 1C andweighed on a Cahn C-33 microbalance (Cahn instru-ments, USA). To investigate the effect of season, thesame experiment was conducted at three differenttimes during the growing season. The first trial wasperformed in July, the second in August and the thirdin November.

All experiments were conduced in a climateroom, 16L:8D photoperiod at 2172 1C and arelative humidity of 70%. Mercury lamps hangingabove the cages provided additional light duringthe photoperiod (40710 mmol photons/m2/s).

Page 5: Temporal changes affect plant chemistry and tritrophic interactions

ARTICLE IN PRESS

Temporal changes affect tritophic interactions 425

GS analysis

In each replicate experiment, leaf samples for GSanalysis were taken from plants damaged byparasitized hosts. The larvae had been feeding onthe plants for 4–8 days and leaves were sampledwhen the larvae moved away from the plant or hadalready pupated. Faeces and pupae were carefullyremoved from the leaves. Leaves were taken fromposition 4–6 on B. oleracea and position 4–7 onS. alba (leaf 1 is the youngest leaf and leaf 4 is thefirst fully developed leaf) and 4–5 leaves weresampled per leaf position. One leaf was taken perplant and all sampled leaves had only moderatetraces of feeding damage. Leaves were cut offusing a razor blade and immediately stored at�80 1C. Leaves were analyzed individually. Sampleswere freeze dried and ground to a fine powder witha mortar and pestle. Aliquots of 50mg for S. albaand 100mg for B. oleracea were weighed in 15mlcentrifuge tubes. GS were extracted and purified asin van Dam, Witjes, and Svatos (2004). GS wereseparated on a reverse phase C-18 column (AlltimaC-18, 3 mm, 150� 4.6mm, Alltech,Deerfield, IL,USA) on HPLC (DIONEX, Sunnyvale, CA, USA) withan acetonitrile water gradient. Detection wasperformed with a DIONEX PDA-100 Photodiode arraydetector set to scan from 200 to 350 nm. Forquantification, sinigrin (Sigma, St. Louis, MO, USA)was used as an external standard. Peaks wereintegrated at 229 nm for which standard responsefactors have been defined (EC, 1990). The differentGS were identified based on their retention timesand UV spectra compared to those of purecompounds (sinigrin, Sigma, St. Louis, MO, USA;glucotropaeolin, sinalbin, and glucobrassicin werekindly provided by M. Reichelt, MPI for ChemicalEcology, Jena, Germany), or compared to acertified oil seed reference (EC Community Bureauof Reference BCR-367R, Fluka, Buchs, Switzerland).For a more detailed description of the method, seevan Dam et al. (2004).

Statistical analysis

Data on adult dry mass and development time ofunparasitized P. xylostella were analyzed using 3-way ANOVA with plant species, sex and replicateand their interactions as factors. For the parasitoiddata, males and females were analysed separatelywith host instar at parasitization, plant speciesand replicate as factors. The Tukey HSD methodwas used for multiple comparisons of means. Toexamine GS content two-way ANOVA were usedwith replicate and leaf position as factors.

Results

GS analysis

The two plant species differed qualitatively andquantitatively in GS content (Fig. 1). In B. oleracea,eight different GS were detected: three alkenyl GS(progoitrin, epiprogoitrin, and sinigrin), four indoleGS (glucobrassicin, 4-hydroxy glucobrassicin,4-methoxy-glucobrassicin, and neoglucobrassicin),and the aromatic glucotropaeolin. The most domi-nant GS in B. oleracea was glucobrassicin. Onaverage, 90% of the GS dry weight mass consisted ofthis compound. In S. alba, four different GS werefound, two alkenyl compounds, i.e. progoitrin andgluconapin, and two aromatic GS, i.e. (gluco)sinal-bin and glucotropaeolin. Sinalbin was the dominantGS representing on average more than 70% of thetotal GS content.

GS concentrations were significantly different forthe three different moments in the season for allGS in both plant species (Table 1). GS contents alsodiffered depending on the position of the leaf onthe plant, with the youngest leaf containing thehighest concentration of GS. In B. oleracea, totalGS levels were 1.5–4.5 fold higher in the 4th leafcompared to the 6th leaf from the top. In S. alba,the total GS concentration was 1.5–2.7 fold higherin the 4th compared to the 7th leaf from the top. InB. oleracea, a significant effect was found for leafposition and GS content for sinigrin, 4- methoxyglucobrassicin and glucobrassicin and in S. alba forgluconapin and sinalbin.

Total GS contents were also different for the twoplant species (Fig. 1). In the July replicate, thetotal GS concentration in leaves at position 4 fromthe top was 40 times higher in S. alba than in B.oleracea. In B. oleracea, the GS concentration inleaves at position 4 from the top was 5.5 timeshigher in November than in July. The change in GSconcentration in S. alba during the season was lessdramatic. The highest concentrations (in leaves atposition 4 from the top) were found in August andNovember and were respectively 1.4 and 1.3 timeshigher than in July.

Effect of host-plant species on the herbivoreP. xylostella

Female P. xylostella readily accepted bothS. alba and B. oleracea as host plants for oviposi-tion. More than 95% of the monitored unparasitizedlarvae developed into adult moths. All factorsincluded in the model had a significant effect onadult biomass (see Table 2). Adult females grew on

Page 6: Temporal changes affect plant chemistry and tritrophic interactions

ARTICLE IN PRESS

2.0

4.0

6.0

8.0

0.0

0.2

0.4

0.6

0.8

SIN

PR

O

EP

RO

GN

A

GB

C

4OH

4MeO

H

NE

O

SN

ALB

TR

OP

TO

TA

L

Glucosinolates

Con

cent

ratio

n (µ

mol

/g D

W)

July

August

November

40.0

60.0

80.0

0.0

5.0

10.0

15.0

20.0

SIN

PR

O

EP

RO

GN

A

GB

C

4OH

4MeO

H

NE

O

SN

ALB

TR

OP

TO

TA

L

Glucosinolates

Con

cent

ratio

n (µ

mol

/g D

W)

A

B

Figure 1. GS concentrations (mean7SE) in B. oleracea(A) and S. alba (B) in leaves sampled from the plant atposition 4 from the top after P. xylostella had completedlarval development (PRO: progoitrin; EPRO: epiprogoi-trin; SIN: sinigrin; 4OH: 4-hydroxy glucobrassicin; 4MeOH:4-methoxy glucobrassicin; TROP: glucotropaeolin; NEO:neoglucobrassicin; GBC; glucobrassicin; GNA: gluconapin;SNALB: sinalbin). Leaves were sampled from plants thathad been used in experiments in July, August andNovember. Four to five plants of each plant species weresampled in each of the three experiments.

R. Gols et al.426

average 1.6 and 1.8 times heavier than males(measured as dry mass) on cabbage and mustard,respectively. The difference in this ratio betweenthe two host plants was consistent over the threereplicates. Female moths were heavier on S. alba(mean dry mass7SE: 1.8270.02mg), than onB. oleracea (1.7170.04mg)(Fig. 2). Moths attainedthe highest biomass in the November replicate onboth plant species (Fig. 2).

There was a significant effect of replicate and amarginal effect of sex on egg-to-adult development

time of the moths (see Table 2). Moths developedfastest in the November replicate (Fig. 3). Femalestook on average 17.070.05 (mean7SE) daysto develop from egg to adult and males took17.270.07 days. There was also a significantinteraction between plant species and replicate,and between plant species and sex (see Table 2).Development of females was faster on S. alba thanon B. oleracea (Fig. 3). The fastest developmentwas found for males on B. oleracea in the Novembertrial (Fig. 3).

Effect of host instar and plant species on theparasitoid D. semiclausum

On average, 83 prepupae or pupae were recov-ered per treatment out of the 100–110 parasitizedlarvae that were put on the plants at the start ofthe experiment (Table 3). About 77% of these pupaeproduced adult D. semiclausum and about 8%produced adult P. xylostella (Table 3). A few deadlarvae were found, but most of the larvae that hadnot pupated could not be retrieved from the plants.

Since D. semiclausum is subject to complemen-tary sex determination (Butcher, Whitfield, &Hubbard, 2000), the number of females emergingfrom parasitized larvae was low, 106 femalescompared to 626 males. Higher sex ratios wereobtained from P. xylostella larvae that wereparasitized as L4, 17.5% females compared to 9%females from L2 larvae. Females grew larger thanmales (ANOVA, F1;734 ¼ 75, Po0:001). Mean drymass of females (7SE) was 0.76170.014mg com-pared to 0.58870.009mg for males. There was nosignificant effect of sex on development time(ANOVA; F1;733 ¼ 0:24, P ¼ 0:63). Wasps took onaverage 15.470.02 days to develop from egg toadult. Since the number of females in some of thetreatments was very low (see Table 3), the effect ofinstar at parasitization, host-plant species andreplicate was analyzed separately for males andfemales.

Males: Wasps were significantly larger and egg-to-adult development time was significantly shorterwhen hosts were parasitized as L4 compared to L2(see Table 2, Fig. 4A and 5A). Wasps also grewlarger and developed faster when the host wasfeeding on S. alba compared to B. oleracea. Thelongest development time and smallest wasps wererecorded for wasps emerging from hosts that wereparasitized as L2 that had fed on B. oleracea in theJuly and August replicate.

Females: As in males, females developed sig-nificantly faster when the hosts were parasitizedas L4 larvae (Table 2, Fig. 5B). The shortest

Page 7: Temporal changes affect plant chemistry and tritrophic interactions

ARTICLE IN PRESS

Table 2. Three-way ANOVA’s on egg-to-adult development time and adult body mass of moth and parasitoid malesand females

Factors Adult dry mass Development time

F-value (df) P-value F-value P-value

Plutella xylostellaTotal model 105 (9,427) o0.001 7.6 (9,925) o0.001Sex 859 (1,427) o0.001 3.7 (1,925) 0.056Plant species 10 (1,427) 0.002 0.01 (1,925) 0.91Replicate 20 (2,427) o0.001 17 (2,925) o0.001Sex� plant species 6.3 (1,427) 0.01 15 (1,925) o0.001Sex� replicate 3.5 (2,427) 0.03 1.4 (2,925) 0.23Plant species� replicate 8.5 (2,427) o0.001 7.1 (2,925) o0.001

Diadegma semiclausum malesTotal model 5.8 (9,616) o0.001 11 (9,615) o0.001Instar 17 (1,616) o0.001 38 (1,615) o0.001Plant species 17 (1,616) o0.001 6.0 (1,615) 0.01Replicate 0.25 (2,616) 0.78 2.8 (2,615) 0.06Instar� plant 3.7 (1,616) 0.054 12 (1,615) o0.001Instar� replicate 2.5 (2,616) 0.08 16 (2,615) o0.001Plant species� replicate 5.5 (2,616) o0.004 7.3 (2,615) o0.001

Diadegma semiclausum femalesTotal model 1.9(9,99) 0.059 3.7 (9,99) o0.001Instar 11 (1,99) 0.002Plant species 3.6 (1,99) 0.06Replicate 3.3 (2,99) 0.04Instar� plant species 2.5 (1,99) 0.11Instar� replicate 3.4 (2,99) 0.04Plant species� replicate 2.9 (2,99) 0.06

Table 1. Two-way ANOVA’s on glucosinolate content with replicate and leaf position as main factors

B. oleracea Replicate Leaf position

Glucosinolate F2,37-value P-value F2,37-value P-value

Progoitrin 6.55 0.004 2.68 0.082Epiprogoitrin 4.91 0.01 1.38 0.26Sinigrin 9.11 o0.001 4.75 0.014-OH glucobrassicin 26.27 o0.001 0.88 0.424-MeOH glucobrassicin 49.64 o0.001 5.75 0.007Glucobrassicin 13.50 o0.001 3.99 0.03Neoglucobrassicin 12.21 o0.001 2.17 0.13Glucotropaeolin 5.07 o0.001 0.88 0.42

Total 15.80 o0.001 4.84 0.01

S.albaGlucosinolate F2,46-value P-value F3,46-value P-value

Progoitrin 4.90 0.011 1.96 0.11Gluconapin 3.36 0.04 3.57 0.01Sinalbin 8.37 o0.001 3.02 0.03Glucotropaeolin 7.21 0.002 1.52 0.21

Total 10.74 o0.001 4.30 0.005

Temporal changes affect tritophic interactions 427

Page 8: Temporal changes affect plant chemistry and tritrophic interactions

ARTICLE IN PRESS

S. alba B. oleracea S. alba B. oleracea

Plant species

0.0

0.5

1.0

1.5

2.0

2.5M

oth

dry

wei

ght (

mg)

July

August

November

Females Males

d (3

0)b (2

6) a (3

6)

c (5

9) b (3

7)

g (3

0)f g

(29)

ef(2

9)

e (2

8)

b (6

6)

eg(4

0)

g (2

7)

Figure 2. Dry weight (mean7SE) of female and maleP. xylostella moths that had been reared on S. alba orB. oleracea. Experiments have been conducted in July,August and November. Bars with the same letter are notsignificantly different (Tukey HSD test for multiplecomparisons among means with a ¼ 0:05). Numbersbetween brackets denote number of individuals (n).

S. alba B. oleracea S. alba B. oleracea

Plant species

15.0

15.5

16.0

16.5

17.0

17.5

18.0

18.5

Mot

h de

velo

pmen

t tim

e (d

ays)

July

August

Novemberdde

bcd

a

d

ab

a

acac

e

a

Females Males

ac

Figure 3. Egg-to-adult development time (mean7SE) offemale and male P. xylostellamoths that had been rearedon S. alba or B. oleracea. Experiments have beenconducted in July, August and November. Bars with thesame letter are not significantly different (Tukey HSD testfor multiple comparisons among means with a ¼ 0:05).Numbers of individuals (n) are the same as in Fig. 1.

Table 3. Recovery (total number of pupae) and fate ofthe pupae of which the larvae had fed on B. oleracea orS. alba and were parasitized by D. semiclausum as second(L2) or fourth instar (L4)

Treatment Pupaetotal #

Moth Waspfemale

Waspmale

Dead

JulyB. oleracea – L2 72 17 5 36 14B .oleracea – L4 78 2 20 44 12S. alba – L2 75 6 8 44 17S. alba – L4 77 0 12 51 14

AugustB. oleracea – L2 83 13 7 56 7B .oleracea – L4 85 1 7 52 25S. alba – L2 61 4 2 47 8S. alba – L4 95 10 12 61 12

NovemberB. oleracea – L2 89 5 7 45 7B .oleracea – L4 96 6 14 71 5S. alba – L2 87 5 8 63 11S. alba – L4 93 9 6 56 22

Pupae developed into P. xylostella moths, D. semiclausum wasps(female or male) or did not emerge (dead).

R. Gols et al.428

development time was found for female waspsdeveloping from hosts that had fed on B. oleraceaand were parasitized as L4 larvae. Biomass was notinfluenced by any of the factors included in themodel (Fig. 4B, Table 2).

Discussion

An overview of the empirical literature revealsmany studies reporting that differences in plantchemistry, particularly secondary compounds can

have a profound effect on the biology, physiologyand ecology of associated consumers, i.e herbivoresand their natural enemies (e.g. Barbosa et al.,1991; Campbell & Duffey, 1981; Francis et al.,2001; Hartmann, 2004; Harvey et al., 2003; Rothet al., 1997; Sznajder & Harvey, 2003). However,the results of most of these studies have beenderived from single experiments that were carriedout at a specific time during the year. In this study,we have shown that, in two closely related plantspecies, there are profound (and species-specific)temporal changes in plant secondary chemistry.These changes occurred even when abiotic condi-tions (temperature, photoperiod) were primarilycontrolled. Moreover, the performance of P. xylos-tella and its parasitoid D. semiclausum also variedbetween replicates, although in both plants it didnot appear to be correlated with GS levels in leaftisues.

In many plant species, such as arable weeds,different populations may germinate and grow atdifferent times of the year, which is true in manyspecies of the Brassicaceae. Our results imply thatunder natural conditions, seasonal shifts in abioticconditions, such as light availability and intensity,as well as temperature, may exert even moredramatic effects on plant ecophysiology and thatthese will have variable effects on higher trophiclevels intimately associated with the plant. Earlierstudies have reported that temporal changes in

Page 9: Temporal changes affect plant chemistry and tritrophic interactions

ARTICLE IN PRESS

S. alba B. oleracea S. alba B. oleracea

S. alba B. oleracea S. alba B. oleracea

0.0

0.2

0.4

0.6

0.8

1.0

0.0

0.2

0.4

0.6

0.8

1.0

July

August

November

L2 L4

bcdg

bcdg

abd

a

abd

bcdg

eg

bcdg ef

g

abf

bcdg

abd

Plant species

Was

p dr

y w

eigh

t (m

g)

L2 L4

Males

Females

a aa

a

aa

a

aa

a

a

a

B

A

Figure 4. Dry weight (mean7SE) of male (A) and female(B) D. semiclausum wasps that had been reared on S. albaor B. oleracea and were parasitized as L2 (first six bars) orL4 (last six bars). Experiments have been conducted inJuly, August and November. Bars with the same letter arenot significantly different (Tukey HSD test for multiplecomparisons among means with a ¼ 0:05). Number ofindividuals (n) in each treatment can be found in Table 1.

S. alba B. oleracea S. alba B. oleracea12.0

13.0

14.0

15.0

16.0

17.0

18.0July

August

November

L2 L4

aab

a a

cbc

aa

aa a

a

S. alba B. oleracea S. alba B. oleracea

Plant species

12.0

13.0

14.0

15.0

16.0

17.0

18.0

Was

p de

velo

pmen

t tim

e (d

ays)

ab abab

ab

a

ab ab

aab

bb

ab

L2 L4

A

B

Females

Males

Figure 5. Egg-to-adult development time (mean7SE) ofmale (A) and female (B) D. semiclausum wasps that hadbeen reared on S. alba or B. oleracea and wereparasitized as L2 (first six bars) or L4 (last six bars).Experiments have been conducted in July, August andNovember. Bars with the same letter are not significantlydifferent (Tukey HSD test for multiple comparisons amongmeans with a ¼ 0:05). Number of individuals (n) in eachtreatment can be found in Table 1.

Temporal changes affect tritophic interactions 429

plant quality affect the performance of bi- ormultivoltine herbivores and their natural enemies.This has been demonstrated in consumers asso-ciated with long-lived plants such as trees wherechanges in the chemistry and toughness of leavesduring the growing season are well documented(Feeny, 1970; Riipi et al., 2002).

Chemical analyses revealed that GS concentra-tions not only varied between plant species but alsoover the different replicates that were carried outat different times during the growing season. Plantsused in the different replicates were of the samesize, but plants in the November replicate took 1week longer to reach the same size as the plantsgrown in July and August. The highest GS concen-trations were found in the November replicate forB. oleracea and in the August and Novemberreplicates for S. alba. Feeny and Rosenberry(1982) reported that GS concentrations in theannual, Brassica nigra, declined during the growing

season. However, in that study the age of theB. nigra plants confounded the results, whereas thepresent study shows that within-plant GS contentsdeclined with leaf age, i.e. older leaves have lowerGS concentrations than younger leaves. Agerbirk,Orgaard, and Nielsen (2003) showed that GSconcentration in natural populations of the cruciferBarbarea vulgaris increase from August to Novem-ber. Furthermore, season and myrosinase activity incultivars of B. oleracea were significantly corre-lated in a study by Charron, Saxton, and Sams(2005) with higher enzyme activity in the fall.

Chemical analyses not only revealed quantitativedifferences but also qualitative differences be-tween GS in leaf tissue of B. oleracea and S. alba.Total GS concentrations were ten to forty timeshigher in S. alba than in B. oleracea, and GScomposition in the two plant species was very

Page 10: Temporal changes affect plant chemistry and tritrophic interactions

ARTICLE IN PRESS

R. Gols et al.430

different, with sinalbin being the dominant GS inS. alba and glucobrassicin in B. oleracea (Franciset al., 2001; Hopkins et al., 1998; Moyes, Collin,Britton, & Raybould, 2000). Cultivated plantspecies often contain lower concentrations ofsecondary plant compounds than their wild rela-tives (Evans, 1993). Artificial selection inB. oleracea is aimed at the production of edibleplant parts, where low levels of bitter-tasting GSare desired (Brussels sprout, cabbage, cauliflower,etc.). By contrast, seeds of S. alba are used as acondiment and usually contain high levels of GS.High levels of GS in seeds are not necessarilycorrelated with high levels of these compounds inother plant tissues (Rosa, 1999), although in thepresent study, much higher foliar concentrations ofGS were found in S. alba than in B. oleracea.Furthermore, GS levels in the cultivated B. oler-acea used here are much lower than GS levels inwild populations of B. oleracea (Moyes et al.,2000). Differences in directional selection inB. oleracea and S. alba may have resulted in theobserved differences in total GS contents.

S. alba appeared to be a qualitatively better hostplant for the development of P. xylostella thanB. oleracea. Male and female moths were differ-entially affected by the host plant upon which theyhad fed. Female moths attained a higher biomasson S. alba than on B. oleracea, whereas males werenot affected by host-plant species. Sexes maydiffer in food requirements and these differencesalready occur in the immature stage when re-sources are acquired for future needs (Slansky,1993). Since there was no difference in develop-ment time between the sexes, female larvae musthave a higher rate of consumption or else are moreefficient in food utilisation than male moths. Astudy by Raps and Vidal (1998) similarly reportedthat plants inoculated with an endophytic fungusaffected the development of female P. xylostellamore than males. Females feeding on endophyte-infested leaves exhibited a reduced conversionefficiency of ingested food and increased theirrelative consumption rates (Raps & Vidal, 1998).

Larval stages of P. xylostella use GS as feedingstimulants, but not all GS are equally effective(Nayar & Thorsteinson, 1963) and an increase in GSconcentration does not always positively correlatewith larval performance (Li et al., 2000). Progoitrinand to a lesser extent glucotropaeolin are strongfeeding stimulants for P. xylostella larvae, whereasgluconapin is toxic at higher concentrations (Nayar& Thorsteinson, 1963). Progoitrin and glucotro-paeolin are found in much higher concentrations inS. alba than in B. oleracea and progoitrin levelsincreased over the growing season in both plant

species. On the other hand gluconapin was onlydetected in S. alba. Thus, performance of theherbivore only partially conformed to the levels ofGS found in damaged leaves. Levels of some GSschange after feeding damage by herbivores (Bart-let, Kiddle, Williams, & Wallsgrove, 1999). In thisstudy, leaf sampling for GS occurred after thelarvae had stopped feeding and therefore reflectsonly the final chemical composition to which theherbivores have been exposed. This could alsoexplain why we did not find a relationship withherbivore performance and plant chemistry. Thepresence of feeding stimulants together withnutrients such as proteins and carbohydrates, aswell as digestibility reducers and deterrents deter-mines food plant quality. It is very likely that notonly GS, but also the primary plant metabolites, aswell as plant compounds that negatively influenceinsect growth varied in the three replicates. Thesemay have also influenced performance of theherbivore in the three replicates.

Host-plant species not only affected herbivoreperformance but also the development of theendoparasitoid, D. semiclausum. Parasitoid malesdeveloped faster and attained the highest biomasswhen attacking hosts feeding on S. alba. However,temporally replicated experiments did not result insignificant effects on parasitoid biomass nor ondevelopment time as had been demonstrated in theherbivore. P. xylostella uses the enzyme, GSsulfatase, which can desulfate a broad range ofGS making them unsuitable substrates for myrosi-nase activity, and thereby circumventing theproduction of toxic hydrolysis products such asnitriles, thiocyanates or isothiocyanates (Ratzka etal., 2002). The host provides all of the resourcesnecessary for parasitoid development and thequality of the host can be affected by secondaryplant compounds ingested by the herbivore (Bar-bosa et al., 1991; Campbell & Duffey, 1981; Harveyet al., 2003; Roth et al., 1997; Sznajder & Harvey,2003). However, detoxification of the GS by thehost may dilute the effect of these compounds onparasitoid growth, although it has been shown thathost-plant species can influence not only the thirdbut the fourth trophic level as well (Harvey et al.,2003). On the other hand, specialised parasitoids,like their hosts, may have also evolved mechanismsto deal with secondary plant compounds.

The effect of host instar at parasitism (L2 or L4)had a strong effect on the development ofD. semiclausum. On both plant species, waspsdeveloped faster and attained the highest biomasswhen the larvae were parasitized as L4. D.semiclausum can successfully develop in all fourlarval stages of P. xylostella, but the offspring sex

Page 11: Temporal changes affect plant chemistry and tritrophic interactions

ARTICLE IN PRESS

Temporal changes affect tritophic interactions 431

ratio (percentage males) decreases when the hostis parasitized as L4 (Yang, Chu, & Talekar, 1993). Inthis study, more females emerged from larvaeparasitized as L4. Many parasitoids lay unferitilizedeggs, which develop into males, in hosts of lowquality (Godfray, 1994). Our results indicate that L4P. xylostella larvae are qualitatively better hostsfor D. semiclausum than L2 larvae. Females of theparasitoid as well as females of the moth are largerthan their male conspecifics. Therefore, it wouldbe interesting to find out when during the devel-opment of the host caterpillar sexual size dimorph-ism occurs and whether an ovipositing wasp is ableto detect sex in the host caterpillar (see alsoGunasena et al., 1989).

A recent study showed that D. semiclausumprefers the odours emitted by S. alba to odoursemitted by B. oleracea (Bukovinszky, Gols, Post-humus, van Lenteren, & Vet, 2005). Moreover,when behaviour of the parasitoid was followed in asemi-field set-up in a greenhouse with both plantspecies and only one Plutella-infested B. oleraceaplant, D. semiclausum females were clearly moreattracted to the S. alba plants even though therewere no hosts present on this plant (Gols et al.,2005). The present study shows that D. semiclau-sum is attracted to the plant species that isqualitatively a better plant for its host, P. xylos-tella, and to a lesser extent for its own develop-ment. It would be interesting to investigatewhether P. xylostella adult females prefer tooviposit and larvae prefer to feed on those plantspecies with the highest nutritional quality. Fieldstudies carried out during the growing seasonshould reveal how changes in plant chemistryaffect oviposition behaviour of adult females.

In summary, our study has revealed that theremay be significant temporal changes in the second-ary chemistry of short-lived plants in whichdifferent populations/cohorts grow at various timesduring the year even under fairly controlledconditions in a greenhouse compartment. In nat-ure, we might expect even more significantseasonal changes in plant chemistry, because thiswould incorporate a range of other abiotic andbiotic selection pressures such as day length, lightintensity, temperature and temporal variation inthe risk of attack from herbivores and pathogens.Importantly, conspecifics of annual plants that growat different times during the year will be exposedto different light intensities and different daylengths, which will very likely affect secondaryplant chemistry much stronger than plants that aregrown at different times during the year undercontrolled greenhouse conditions. Different gen-erations of multivoltine insect herbivores may

therefore feed on individual plants whose chem-istry varies dramatically, and by association theirparasitoids may also experience considerable sea-son-related differences in host-plant quality, whichmay affect their development. Longer-term studiesare thus required in order to better understand howpotentially dynamic changes in plant quality affecta range of multitrophic interactions and commu-nity-level processes.

References

Agerbirk, N., Orgaard, M., & Nielsen, J. K. (2003).Glucosinolates, flea beetle resistance, and leaf pub-escence as taxonomic characters in the genus Barbar-ea (Brassicaceae). Phytochemistry, 64, 1177.

Agrawal, A. A. (1999). Induced plant defense: Evolutionof induction and adaptive phenotypic plasticity. In A.A. Agrawal, S. Tuzun, & E. Bent (Eds.), Induced plantdefenses against pathogens and herbivores. Biochems-try, ecology and agriculture (pp. 251–268). St. Paul,MN: APS Press.

Agrawal, A. A., & Kurashige, N. S. (2003). A role forisothiocyanates in plant resistance against the specia-list herbivore Pieris rapae. Journal of ChemicalEcology, 29, 1403–1415.

Barbosa, P., Gross, P., & Kemper, J. (1991). Influence ofplant allelochemicals on the tobacco hornworm andits parasitoid, Cotesia congregata. Ecology, 72,1567–1575.

Bartlet, E., Kiddle, G., Williams, I., & Wallsgrove, R.(1999). Wound-induced increases in the glucosinolatecontent of oilseed rape and their effect on subsequentherbivory by a crucifer specialist. Entomologia Ex-perimentalis et Applicata, 91, 163–167.

Bukovinszky, T., Gols, R., Posthumus, M. A., vanLenteren, J. C., & Vet, L. E. M. (2005). Variation inplant volatiles and the attraction of the parasitoidDiadegma semiclausum (Hellen). Journal of ChemicalEcology, 31, 461–480.

Butcher, R. D. J., Whitfield, W. G. F., & Hubbard, S. F.(2000). Complementary sex determination in thegenus Diadegma (Hymenoptera: Ichneumonidae).Journal of Evolutionary Biology, 13, 593–606.

Campbell, B. C., & Duffey, S. S. (1981). Alleviation of a-tomatine-induced toxicity to the parasitoid, Hyposo-ter exiguae, by phytosterols in the diet of the host,Heliothis zea. Journal of Chemical Ecology, 7,927–946.

Charron, C. S., Saxton, A. M., & Sams, C. E. (2005).Relationship of climate and genotype to seasonalvariation in the glucosinolate-myrosinase system. II.Myrosinase activity in ten cultivars of Brassicaoleracea grown in fall and spring seasons. Journal ofthe Science of Food and Agriculture, 85, 682–690.

Dussourd, D. E. (1993). Foraging with finesse: Caterpillaradaptations for circumventing plant defenses. In N. E.Stamp, & T. M. Casey (Eds.), Caterpillars: Ecologicaland evolutionary constraints on foraging (pp. 92–131).

Page 12: Temporal changes affect plant chemistry and tritrophic interactions

ARTICLE IN PRESS

R. Gols et al.432

EC. (1990). Oil seeds – Determination of glucosinolateshigh perfomance liquid chromatography. Official Jour-nal of the European Communities, L 170/28, AnnexVIII, 03.07.27–34.

Evans, L. T. (1993). Crop evolution, adaptation and yield.Cambridge: Cambridge University Press.

Fahey, J. W., Zalcmann, A. T., & Talalay, P. (2002). Thechemical diversity and distribution of glucosinolatesand isothiocyanates among plants. Phytochemistry,59, 237.

Feeny, P. (1970). Seasonal changes in oak leaf tannins andnutrients as a cause of spring feeding by winter mothcaterpillars. Ecology, 51 565.

Feeny, P., & Rosenberry, L. (1982). Seasonal variation inthe glucosinolate content of North-American Brassicanigra and Dentaria species. Biochemical Systematicsand Ecology, 10, 23–32.

Fox, L. R., Kester, K. M., & Eisenbach, J. (1996). Directand indirect responses of parasitoids to plants: Sexratio, plant quality and herbivore diet breadth.Entomologia Experimentalis et Applicata, 80,289–292.

Francis, F., Lognay, G., Wathelet, J. P., & Haubruge, E.(2001). Effects of allelochemicals from first (Brassica-ceae) and second (Myzus persicae and Brevicorynebrassicae) trophic levels on Adalia bipunctata. Journalof Chemical Ecology, 27, 243–256.

Godfray, H. C. J. (1994). Parasitoids. Behavioral andevolutionary ecology. Princeton, New Jersey: Prince-ton University Press.

Gols, R., Bukovinszky, T., Hemerik, L., Harvey, J. A., VanLenteren, J. C., & Vet, L. E. M. (2005). Reducedforaging efficiency of a parasitoid under habitatcomplexity: Implications for population stability andspecies co-existence. Journal of Animal Ecology, 74,1059–1068.

Gouinguene, S. P., & Turlings, T. C. J. (2002). The effectsof abiotic factors on induced volatile emissions in cornplants. Plant Physiology, 129, 1296–1307.

Gunasena, G. H., Vinson, S. B., & Williams, H. J. (1989).Interrelationships between growth of Heliothis vires-cens (Lepidoptera: Noctuidae) and that of its para-sitoid, Campoletis sonorensis (Hymenoptera,Ichneumonidae). Annals of the Entomological Societyof America, 82, 187–191.

Gupta, P. D., & Thorsteinson, A. J. (1960a). Food plantrelationships of the diamondback moth (Plutellamaculipennis (Curt.)). I. Gustation and olfaction inrelation to botanical specificity of larva. EntomologiaExperimentalis et Applicata, 3, 241–250.

Gupta, P. D., & Thorsteinson, A. J. (1960b). Food plantrelationships of the diamondback moth (Plutellamaculipennis (Curt.)). II. Sensory regulation andoviposition of the adult female. Entomologia Experi-mentalis et Applicata, 3, 305–314.

Hartmann, T. (2004). Plant-derived secondary metabo-lites as defensive chemicals in herbivorous insects: Acase study in chemical ecology. Planta, 219, 1–4.

Harvey, J. A., van Dam, N. M., & Gols, R. (2003).Interactions over four trophic levels: Foodplant

quality affects development of a hyperparasitoidas mediated through a herbivore and itsprimary parasitoid. Journal of Animal Ecology, 72,520–531.

Hopkins, R. J., Ekbom, B., & Henkow, L. (1998).Glucosinolate content and susceptibility for insectattack of three populations of Sinapis alba. Journal ofChemical Ecology, 24, 1203–1216.

Karban, R., & Baldwin, I. T. (1997). Induced responses toherbivory. Chicago: University of Chicago Press.

Karban, R., & Myers, J. H. (1989). Induced plantresponses to herbivory. Annual Review of Ecologyand Systematics, 20, 331–348.

Li, Q., Eigenbrode, S. D., Stringham, G. R., & Thiagar-ajah, M. R. (2000). Feeding and growth of Plutellaxylostella and Spodoptera eridania on Brassica junceawith varying glucosinolate concentrations and myr-osinase activities. Journal of Chemical Ecology, 26,2401–2419.

Mithen, R. (2001). Glucosinolates – biochemistry, genet-ics and biological activity. Plant Growth Regulation,34, 91–103.

Moyes, C. L., Collin, H. A., Britton, G., & Raybould, A. E.(2000). Glucosinolates and differential herbivory inwild populations of Brassica oleracea. Journal ofChemical Ecology, 26, 2625–2641.

Muller, C., Agerbirk, N., Olsen, C. E., Boeve, J. L.,Schaffner, U., & Brakefield, P. M. (2001). Sequestra-tion of host plant glucosinolates in the defensivehemolymph of the sawfly Athalia rosae. Journal ofChemical Ecology, 27, 2505–2516.

Naumann, C., Hartmann, T., & Ober, D. (2002). Evolu-tionary recruitment of a flavin-dependent monoox-ygenase for the detoxification of host plant-acquiredpyrrolizidine alkaloid-def ended arctiid alkaloids inthe moth Tyria jacobaleae. Proceedings of theNational Academy of Sciences of the United Statesof America, 99, 6085–6090.

Nayar, J. K., & Thorsteinson, A. J. (1963). Furtherinvestigations into the chemical basis of insect-hostplant relationships in an oligophagous insect, PlutellaMaculipennis (Curtis) (Lepidoptera – Plutellidae).Canadian Journal of Zoology, 41, 923.

Ohsaki, N., & Sato, Y. (1999). The role of parasitoids inevolution of habitat and larval food plant preferenceby three Pieris butterflies. Researches on PopulationEcology, 41, 107–119.

Pivnick, K. A., Jarvis, B. J., & Slater, G. P. (1994).Identification of olfactory cues used in host-plantfinding by diamondback moth, Plutella xylostella(Lepidoptera, Plutellidae). Journal of Chemical Ecol-ogy, 20, 1407–1427.

Raps, A., & Vidal, S. (1998). Indirect effects of anunspecialized endophytic fungus on specializedplant–herbivorous insect interactions. Oecologia,114, 541–547.

Rask, L., Andreasson, E., Ekbom, B., Eriksson, S.,Pontoppidan, B., & Meijer, J. (2000). Myrosinase:Gene family evolution and herbivore defense inBrassicaceae. Plant Molecular Biology, 42, 93–113.

Page 13: Temporal changes affect plant chemistry and tritrophic interactions

ARTICLE IN PRESS

Temporal changes affect tritophic interactions 433

Ratzka, A., Vogel, H., Kliebenstein, D. J., Mitchell-Olds,T., & Kroymann, J. (2002). Disarming the mustard oilbomb. Proceedings of the National Academy ofSciences of the United States of America, 99,11223–11228.

Reed, D. W., Pivnick, K. A., & Underhill, E. W. (1989).Identification of chemical oviposition stimulants forthe diamondback moth, Plutella xylostella, present inthree species of Brassicaceae. Entomologia Experi-mentalis et Applicata, 53, 277–286.

Renwick, J. A. A. (2002). The chemical world ofcrucivores: Lures, treats and traps. EntomologiaExperimentalis et Applicata, 104, 35–42.

Renwick, J. A. A., & Lopez, K. (1999). Experience-basedfood consumption by larvae of Pieris rapae: Addictionto glucosinolates? Entomologia Experimentaliset Applicata, 91, 51–58.

Riipi, M., Ossipov, V., Lempa, K., Haukioja, E., Koricheva,J., Ossipova, S., et al. (2002). Seasonal changes inbirch leaf chemistry: Are there trade-offs betweenleaf growth, and accumulation of phenolics? Oecolo-gia, 130, 380–390.

Rosa, E. A. S. (1999). Chemical composition. In C. Gomez-Campo (Ed.), Developments in plant genetics andbreeding, 4. Biology of Brassica coenospecies(pp. 315–357). Amsterdam: Elsevier Science B.V.

Roth, S., Knorr, C., & Lindroth, R. L. (1997). Dietaryphenolics affects performance of the gypsy moth(Lepidoptera: Lymantriidae) and its parasitoid Cotesiamelanoscela (Hymenoptera: Braconidae). Environ-mental Entomology, 26, 668–671.

Schoonhoven, L. M., van Loon, J. J. A., & Dicke, M.(2005). Insect–plant biology (2nd ed). Oxford: OxfordUniversity Press.

Slansky, F. (1993). Nutritional ecology: the fundamentalquest for nutrients. In N. E. Stamp, & T. M. Casey(Eds.), Caterpillars: Ecological and evolutionary con-straints on foraging (pp. 29–91).

Steppuhn, A., Gase, K., Krock, B., Halitschke, R., &Baldwin, I. T. (2004). Nicotine’s defensive function innature. Plos Biology, 2, 1074–1080.

Stout, M. J., Brovont, R. A., & Duffey, S. S. (1998). Effectof nitrogen availability on expression of constitutive andinducible chemical defenses in tomato, Lycopersiconesculentum. Journal of Chemical Ecology, 24, 945–963.

Stowe, K. A. (1998). Realized defense of artificiallyselected lines of Brassica rapa: Effects of quantitativegenetic variation in foliar glucosinolate concentra-tion. Environmental Entomology, 27, 1166–1174.

Sznajder, B., & Harvey, J. A. (2003). Second and thirdtrophic level effects of differences in plant speciesreflect dietary specialisation of herbivores and theirendoparasitoids. Entomologia Experimentalis et Ap-plicata, 109, 73–82.

Takabayashi, J., Dicke, M., & Posthumus, M. A. (1994).Volatile herbivore-induced terpenoids in plant miteinteractions – variation caused by biotic and abioticfactors. Journal of Chemical Ecology, 20, 1329–1354.

Traw, M. B., & Dawson, T. E. (2002). Reduced perfor-mance of two specialist herbivores (Lepidoptera:Pieridae, Coleoptera: Chrysomelidae) on new leavesof damaged black mustard plants. EnvironmentalEntomology, 31, 714–722.

van Dam, N. M., Hadwich, K., & Baldwin, I. T. (2000).Induced responses in Nicotiana attenuata affectbehavior and growth of the specialist herbivoreManduca sexta. Oecologia, 122, 371–379.

van Dam, N. M., Witjes, L., & Svatos, A. (2004).Interactions between aboveground and belowgroundinduction of glucosinolates in two wild Brassicaspecies. New Phytologist, 161, 801–810.

van Loon, J. J. A., & Schoonhoven, L. M. (1999). Specialistdeterrent chemoreceptors enable Pieris caterpillars todiscriminate between chemically different deterrents.Entomologia Experimentalis et Applicata, 91, 29–35.

Yang, J. C., Chu, Y. I., & Talekar, N. S. (1993). Biologicalstudies of Diadegma semiclausum (Hym, Ichneumoni-dae), a parasite of diamondback moth. Entomophaga,38, 579–586.

Zangerl, A. R., & Berenbaum, M. R. (1993). Plantchemistry, insect adaptations to plant chemistry, andhost plant utilization patterns. Ecology, 74, 47–54.