Interactions Among Three Trophic Levels: Influence of Plants on Interactions Between Insect Herbivores and Natural Enemies Author(s):

Peter W. Price, Carl E. Bouton, Paul Gross, Bruce A. McPheron, John N. Thompson, Arthur E. Weis Source: Annual Review of Ecology and Systematics, Vol. 11 (1980), pp. 41-65 Published by: Annual Reviews Stable URL: http://www.jstor.org/stable/2096902 Accessed: 30/01/2009 18:22
Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in the JSTOR archive only for your personal, non-commercial use. Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at http://www.jstor.org/action/showPublisher?publisherCode=annrevs. Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed page of such transmission. JSTOR is a not-for-profit organization founded in 1995 to build trusted digital archives for scholarship. We work with the scholarly community to preserve their work and the materials they rely upon, and to build a common research platform that promotes the discovery and use of these resources. For more information about JSTOR, please contact support@jstor.org.

Annual Reviews is collaborating with JSTOR to digitize, preserve and extend access to Annual Review of Ecology and Systematics.

http://www.jstor.org

Ann.Rev.Ecol Syst 1980.11:41-65 ? 1980 byAnnualReviews Copyrght Ina All rghts reserved

INTERACTIONS AMONG THREE TROPHIC LEVELS: Influence of Plants on Interactions Between Insect Herbivores and Natural Enemies
Peter W Price1, Carl E. Bouton, Paul Gross, Bruce A. McPheron, John N. Thompson 2, and Arthur E. Weis
of Entomology, of Illinois,Urbana,Illinois61801 Department University

+4171

INTRODUCTION
In his recentreviewof the developing theoryof insect-plant interactions, Gilbert(52) identified fourmajorthrustsin research: coevoluinsect-plant tion, host plants as islands,plant apparencyand chemicaldefense,and resourcepredictability versusevolutionary strategiesof insects.From his reviewit is evidentthat, with a few exceptions (23, 45, 53, 64), developing theory is addressing primarilya two trophic level system. In reality,of course,all terrestrial communities basedon living plantsare composedof at least three interactingtrophic levels: plants, herbivores,and natural enemiesof herbivores. We arguethat theoryon insect-plant interactions cannotprogress realisof the thirdtrophiclevel.A closerlook at the ticallywithoutconsideration mechanisms of interactions reveals a paradox, andplantshavemanyeffects, directand indirect,positiveand negative,not only on herbivores but also on the enemiesof herbivores. as The thirdtrophiclevel mustbe considered part of a plant'sbatteryof defensesagainstherbivores.
'Present address: Museum of Northern Arizona86001 Arizona,Flagstaff, 2Present of BotanyandZoology,Washington StateUniversity, Pulladdress: Departments 99164 man,Washington 41

0066-4162/80/1120-0041$01.00

42

et al PRICE

insect We focus on insect herbivoresand their enemies, particularly Much is relevantto and parasiticwaspsand flies (parasitoids). predators is richerthanwe can andtheirenemies,andthe literature otherherbivores indicatein a reviewof this size (see 93 for anotherview). To clarifyand only for enemies will be employed discussion, the term"enemy" abbreviate and parasites. parasitoids, predators, of herbivores:

THEORYOF INTERACTIONS
andenemiesof herbivores amongplants,herbivores, The interrelationships by severalareas of theoreticalecology that have not are encompassed usuallybeenviewedin the samecontext:theoryof plantchemicaldefense, and food web modelsof insect herbivores, generaldynamicalpopulation and reinforcfromthis body of theoryare overlapping theory.Predictions begunto be testedin the contextof thethirdtrophic ingbuttheyhavebarely level. by Feeny(44, 45) and The theoryof plantchemicaldefenseas developed of the aboutthe efficacy Rhoades& Cates(99) permitscertainpredictions by the life historyof the plant. influenced enemiesas ultimately herbivores' and patchily Plantspeciesin the earlystagesof successionare short-lived (45). (99) and "unapparent" and so are very "unpredictable" distributed is importhat earlycolonization life meansfor herbivores This ephemeral to discovera plant(or patchof plants)can be tant.The earliestherbivores expected,by chance alone, to arrivefrom nearbyplants, but in patchy species.Therefore, such sourceplantsare usuallyof different distributions will of a givenplantspeciesevolves,individuals chemistry as the defensive Cates (99) be favoredthat producechemicals[the "toxins"of Rhoades& from different of Feeny (44, 45)] increasingly defenses" and "qualitative will evolvethe at leastsomeherbivores thoseof associated plants.Probably chemical capacityto detoxifysuch compounds,but a high interspecific diversityshould pose a formidablebarrierto the evolutionof extreme a plantwitha uniquechemis(seealso 49). Therefore, herbivore polyphagy defended frommostinsects;it will oftenescapeeven try will be chemically on tendto specialize (andtherefore insectsthatcan detoxifyits compounds it will oftenoccurat somedistancefromits nearestconspecifics. it) because findandovipositon theseplants,theiroffspring shouldspecialists However, of the specifictoxins of the concentration will developrapidly-regardless find them.In addition, beforeenemiescan (45)-and maybe ableto mature in their guts), least they will containthe chemicalsof the host plant (at also 23, 52). enemies (see which are likely to have directtoxic effectson situathe opposite Feeny(44, 45) and Rhoades& Cates(99) arguethat the dominant least (at tion prevailsin the defensivestrategyof plants

INTERACTIONS AMONG THREETROPHICLEVELS

43

species)in late-successional, temperate systems.Theseplantsarelong-lived andtendto growin standsof low diversity; they arethusvery"predictable" (99)and "apparent" (45) resources. Mostplantsin sucha systemwouldnot be able to escape the averageherbivorespecies for more than a small fractionof theirlifespan. Therefore, plantsthat produceonly toxinswould be extremely vulnerable to at least a smallgroupof specialists, and chemically distincttoxinswouldseem to conferlittle advantage (but see 36). In accordancewith these factors,naturalselection has apparently favored mechanisms defensive less lethal,cannotbe cirthat, althoughpotentially cumvented withoutsubstantial cost to the herbivore. The commondefento herbivores sive strategyseemsto involvea reductionof the availability of essentialnutrients.This has been achievedby a heavy investmentin of Rhoades& Cates(99) certaincompounds-the "digestibility reducers" or the "quantitative defenses" of Feeny(44, 45)-including tannins(42, 43), resins(44, 99), proteolytic enzymeinhibitors (107), and silica (25). These compoundscan interferewith digestiveefficiencyand prolong the herbivores'development. The apparentlylower chemical diversityof such systems, when coupled with the fact that several digestibilityreducing mechanisms have a similarmode of action (they interactwith proteins
within the gut), should result in greater herbivore polyphagy in late successional habitats. Futuyma (49) has shown such increased polyphagy for

All of theseeffectson the herbivores Lepidoptera. can serveto enhancethe of enemiesin comparison effectiveness to early successional systems:(a) Greater herbivore on plantsthat arealreadymoreapparent and polyphagy predictable shouldresultin herbivores that are also apparent and predictable. (b) Prolonged development time results in longer exposure of the more vulnerable, immature stages to enemies (45). (c) Herbivores that feed on

nontoxicfood cannotsequester chemicalsfor defense(45). The theorybasedon generalpopulation modelsof herbivores and their enemiesdevelopssimilarconclusions. Southwood (119) deviseda synoptic model of population dynamicsincorporating populationgrowth,population density, and habitatstability.Elaborated by Southwood& Comins (121) and Southwood(120), the model predictedthat enemies are less importantas a controllinginfluenceon the populationat low levels of habitatstability(their "r-selected" habitatstability)than at higherlevels of habitatstability.Organisms at the low end of the habitat-stability axis are likely to escape from enemies in space and time (121). Supportive evidencefor the population-level of this modelcomesfromthe predictions literature on biological control.Manyauthors(e.g. 120, 140)haveanalyzed the relativesuccessof biological controlattemptsin agroecosystems versus orchardand forestsettings.The trendis for more successfulintroduction in orchard-forest andestablishment of enemies habitats thanin agroecosys-

44

PRICE et al

tems.Agricultural systemsaresimilarin manyrespectsto an early-successionalhabitat: Disturbances createconditions for the continued unsuitable coexistenceof plant, herbivore, and enemy (140), and colonizationis an importantaspect of the system.Orchard-forest systems represent conditions closer to a late-successional habitat.Long-term interactions among the threetrophiclevelsaremorelikelyto occursimplybecausethe role of disturbance and recolonization is minimized. Food web stabilityas a functionof food chain lengthand the degreeof of organisms has beentheoretically specialization investigated by May (76, 77), MaynardSmith (78), Beddington& Hammond(7), and Pimm & Lawton(88, 89). These authorspredictthat, in general,stabilityshould decreaseas the numberof connectinglinks in the food web increases. and their enemies,such as may be found on toxic Specialized herbivores plantsin early succession,producefood webs of low connectance. More andtheirenemies, generalized herbivores as mayoccuron late-successional plants, show a decreasedstabilityas a result of the greaternumberof interconnecting links. These predictedtrendsmay, however,be counterbalancedby the relativeinstabilityof early-successional resources.Any realisticprediction of degreeof stabilitydependson the relativebalanceof at eachtrophiclevel and on the stabilityof the environment specialization in which the food web exists.

PROPERTIES OF INDIVIDUAL PLANTS

Traitsof individual betweenherbivores plantsmaymodifyinteractions and their enemiesby operating the enemy,or both. directlyon the herbivore, Thesetraitsmay be eitherchemical(such as toxins,digestibility-reducers, andnutrient or physical andtissuetoughness). balance) (suchas pubescence

Effects That OperateDirectly on Herbivores

The opportunity to influence and enemies interactions betweenherbivores would not be greatif plantdefensivetraitsresultedin rapiddeathof the herbivore. Enemieswouldhaveno role to play. In addition,the herbivores themselveswould probablysoon evolve either an effectivedetoxification ability[so thatthe traitwouldhavea negligible effect(45)] or an avoidance response. For thesereasons,the less potentdigestibility-reducers probably havefarmoresignificant thando toxins.Digestidirecteffectson herbivores can exertsublethal effectsin threemajorways:by impairing bility-reducers growth,loweringresistance to disease,and reducingfecundity. of herbivore Impairment growthcan resultin a prolonged development time. This may have important for interaction betweenthe consequences second and third trophiclevels. That digestibility-reducers can prolong

INTERACTIONS AMONG THREETROPHICLEVELS

45

development time is suggested by the fact that lepidopteran larvae feeding on mature, high-tannin (up to 5.5% of leaf dry weight) oak leaves grow much more slowly than those on young, low-tannin leaves (43). Some cannot even reach adulthood in a single season. More precise evidence comes from artificial diet studies by Chan et al (27) with the lepidopteran Heliothis virescens,a pest of cotton. They found that the presence of only 0.3% (dry weight) of condensed tannin (extracted from cotton) in the diet increased development time by 21 %, even though the insects were switched to a normal diet before half of their larval life had been completed. Texas 254, a cotton variety resistant to this insect, contains 0.7% condensed tannin. Prolonged development time should increase the herbivore'srisk of attack by enemies during its vulnerable immature stages (45). This prediction is supported by a review of the literatureon parasitoids on Lepidoptera in early versus late-successional habitats (B. A. McPheron and P. Gross, unpublished), which indicates that in general significantly more parasitoid species colonized the latter: a mean of 2.8 parasitoid species on earlysuccessional Lepidoptera and 6.4 species on those in late succession. Especially important may be the improved probability of temporal coincidence of insect herbivores with specialized parasitoids, which usually attack only particular instars of their host. Host populations that are developmentally well-synchronized may only include individuals of these suitable instars for short periods (56) unless development is prolonged. Digestibility-reducers can also decrease herbivores' resistance to pathogens. Chan et al (27) found that only 63% of Heliothis virescens larvae survived to pupation on a high-tannin diet compared to 81% on one that was tannin-free. The role of pathogens was not investigated but, since tannins are probablynot directly toxic (99), the decreased survival may well be due to stress that weakens herbivore resistance to such organisms (124). Stubblebine& Langenheim(127) rearedthe armyworm, Spodopteraexigua, on artificial diets with and without a leaf resin extracted from the legume, Hymenaea courbaril. One experiment was performed with larvae from a single egg mass that had been infected (inadvertantly) with virus. Viral mortality ranged from 0% among larvae on the resin-freediet to over 50% among those on the diet containing 1.6% resin (dry weight). Digestibility-reducers may also affect herbivore-enemy interactions by causing decreased herbivore body size and hence fecundity [see (24) for correlations between body size and fecundity]. For example, when Feeny (42) reared larvae of the winter moth, Operophterabrumata, on artificial diets containing tannin concentrations equivalent to that of Septemberoak leaves, the resultant pupae were significantly smaller than the control insects (22.4 mg vs. 30.8 mg). If smaller size results in lower fecundity, this will lower the rate of population increase. This greatly influences the ability

46

PRICE et al

of enemiesto control a local herbivorepopulation[(139, 152) and see below]. Influence on enemyeffectiveness throughdirectinfluence on herbivores is not limitedto digestibility-reducing compounds; any plantcharacter that can affect growth,resistanceto disease,or fecunditycan have such an influence (e.g. 137).Variation in plantnutrientcontentis probably importantin this regard. Whilesuchvariation probably seldominvolvescomplete presence or absence of essential nutrients (48), changesin theirproportions appearto be commonplace (62). Effectsthat are less than lethal should, therefore,also be commonplace. Negativeeffectsof imbalanced diets on insectsurvivorship, growth,and development have, not surprisingly, been in controlledartificialdiets for a numberof insects and demonstrated nutrients(31). Thatsuchmechanisms in natureis stronglysuggested operate by numerous observed correlations betweeninsectperformance andlevelsof various in plants.Mostthoroughly nutrients studiedis the roleof nitrogen, recently reviewedby MeNeill & Southwood(79). Most studies involve temporal fluctuations of nitrogenous compounds withinsingleplants.Suchstudiesno doubt reveal importantecologicalinfluencesupon interactionsbetween herbivoresand enemies,but they provideonly limited insight into the potentialevolutionary importancefor plants of a geneticallycontrolled variation of nitrogenous Oneexceptionis the resistance compounds. of rice to the riceplanthopper andthe riceleafhopper. Resistant varieties had low levels of aminoacids,especiallyasparagine (22), which resultedin significantreductions in leafhopper rateof weightgainand fecundity (28). Survivorship was also significantlyreduced, probably owing to increased to pathogens. susceptibility Slansky& Feeny (115) also investigated a system involvingvariation in levelsof nitrogenous betweendiffercompounds ent speciesof crucifers. In this case,however,the lepidopteran Pierisrapae grewjust as fast on low- as on high-nitrogen plantsbecauseit consumed more of the formerand utilizedthe nitrogenwithin that food more effias Slansky& Feeny (115) point out, low nitrogen ciently. Nevertheless, levelsmaystillincrease herbivore to enemies.P. rapaeactively vulnerability feedsfor only a smallfraction of everyday andspendsthe remainder of the time resting and digestingfood in a more cryptic position. In order to increaseconsumptionof low-nitrogenplants, larvae must spend longer periodsmore exposedto enemies. Othervariable factorsmayproduce relevant sublethal effectsin herbivorous insects.The importance of leaf watercontenthas been demonstrated for growthof the cecropia moth,Hyalophora cecropia, (108),andfor other mothsand butterflies can also (109). Physicalfactors,such as pubescence, be important. Wellso(149) foundthat cerealleaf beetles,Oulemamelano-

AMONG THREETROPHICLEVELS INTERACTIONS

47

time, loweradultweight,and lower increased development pus, exhibited when rearedon small grain cultivarswith greatertrichome survivorship before densities.Smalllarvaehad to bite individualtrichomesrepeatedly gainingaccess to the nutritiousadaxialcells. Later instarsconsumetriand capable indigestible chomeswhole,but theyaresiliceousandtherefore the midgut.Thispermits entryinto the hemocoelof "potential of lacerating anaerounderthe relatively that are nonpathogenic pathogens"-bacteria bic conditionsof the gut but lethal in the body cavity (19). Finally,even the so-calledtoxins(99) can havethese sortsof sublethaleffects.Erickson Papilio & Feeny(40) foundreducedgrowthratesof the blackswallowtail, & polyxenes,on dietsto whichsinigrinhad beenadded.Also, Stubblebine susceptibility (127)foundthat legumeresins,whichincreased Langenheim S. exigua, to viralinfection,may actuallybetterfit the of the armyworm, definitionof toxins-small, lipid-solublemolecules(99). Resins had no insects, so their direct effectsare amongvirus-free effecton survivorship truly sublethal. of sublethaleffectshave thus far been porThe three majorcategories betweenherbivores and interactions trayedas thoughthey couldinfluence time of that operatewithin the generation enemieseitherby mechanisms of impaired susceptibilthe herbivore growthand increased (the categories (the reduced actingacrossgenerations ity to disease)or elseby mechanisms susceptibility However, impaired growthandincreased category). fecundity For example,an impaired growthrate can also operateacrossgenerations. reducesa population's can prolonggeneration time, which dramatically (29). And, since rateof increaseis also a function intrinsicrateof increase to susceptibility of age-specific mortalitydue to an increased survivorship, diseaseagentswill, of course,also reducethe rate of increase.A rate of thatis specific to an individual plant(or smallpatchof plants)may increase be especiallypertinentto small, relativelysessile herbivoreslike mites, which spend more than a single generationon a aphids,and whiteffies, singleplant. The reductionof rates of increaseby plant factors can be of central popuin determining the abilityof enemiesto controlherbivore importance VanEmden(139) demonstrated the potentialimporlations.For example, tance of variablerates of herbivoreincreaseon differentplants, in the fixed contextof predation, equationmodel.A particular usinga difference insect that had the predationrate was unableto controla phytophagous of multiplying1.2-foldeach day on a potentialin the absenceof predators givenhost plant.However,controlwas achievedwhenthe rateof increase since wasonlyslightlyless:1.15.Of course,this is a grossoversimplification Neverrate ignoresfunctionaland numerical a fixedpredation responses. theless, it does illustratethe potentialfor at least temporarynumerical

48

PRICEet al

escapebeforea predator's own numerical response is manifested (assuming the predatorhas a long generation time). VanEmden's modelhasbeentestedwith experimental systemsinvolving resistant cultivars on whichparticular herbivorous insectsrealizedreduced ratesof increase. Starkset al (122) testedthe effectof a parasitoid, Lysiphlebustestaceipes, on population growthof the aphid,Schizaphis graminum, on resistant andsusceptible varieties of barley.Absolutemortality attributable to parasitoids was similaron the two varieties.However,since there werefewertotal insectson the resistantvariety,the ratesof parasitism on it were much higher.These higherratesenabledthe parasitoids to check aphidpopulation growthon the resistant varietyat all initiallevelsof aphid whereas infestation, they wereonly ableto do so on the susceptible variety level(three).Withinoculums of six or twelveaphids at the lowestinoculum at an exponential rateat per susceptible plant,aphidswerestill increasing the endof fourweeks.Theseresultsareconsistent withthoseof Wyatt(155, 156),who studiedpopulation growthof the aphid,Myzuspersicae, on four cultivarswith and withoutthe parasitoid, chrysanthemum Aphidiusmatricariae. betweenpredation and Finally,Dodd (33) studiedthe interaction on different plant resistancein the cabbageaphid, Brevicoryne brassicae, varietiesof brusselssprouts.The potentialgrowthrateof aphidswas only slightlyless on the variety"EarlyHalf Tall" than on "WinterHarvest." However,when aphidpopulations declinedduringinclementconditions, the residualmortality to predation was muchgreateron "Early attributed Half Tall." Since predatordensitieswere not controlled,however,it is impossibleto determinewhether each predatorwas more effectiveon to "EarlyHalfTall"or whetherthis planttype was simplymoreattractive certainpredators.

Directlyon Enemies Effects That Operate

Theactivities of natural enemiescanbe influenced directlyin eitherpositive or negative of theirhosts'foodplants.These of properties waysby a variety characteristics havefar-reaching effectson resourceutilizationpatternsin herbivores. necThe availability of secretednutrients,particularly the presenceand abundance of not only mutualists tar, greatlyinfluences with the plant-particularlyants (9), but also otherpredators and parasithe toids (51, 53, 116). The adaptivefinesseof the plant in manipulating protective thirdtrophiclevelis illustrated by Tilman's (135) demonstration maximum secretionfromextrafloral nectaries that black cherrymaintains at the time when ants are most able to prey upon the majorherbivore, of the literaMalacosoma americanum. Bentley's(9) reviewcoversm.uch
ATTRACTANTS

INTERACTIONS AMONG THREE TROPHIC LEVELS

49

ture on plant-antmutualism,althoughshe omits mentionof extrafloral nectarieson galls. The existenceof such nectarieswas noticedby Theo(10) lists 16 Europeanspeciesof cynipidwasps phrastus(133). Bequaert galls;manyothersexist in North America(146-148). secreting producing a hostof insects,includby thesegallsattracts The copiousnectarproduced review(10) suggests, Bequaert's parasitoids. ing, no doubt,the gall-makers' thatantsmayoftenexcludeothernectarfeeders(e.g.parasitoids), however, therebyprotectingthe gall-maker. The productionof floralnutrientsallows a loose mutualismto form was The tarnished plantbug, Lyguslineolaris, between plantsandenemies. pallipes on weedy speciesof hardlyattackedby the braconidLeiophron commonly and Solidago,but parasitism Daucus,Amaranthus, Oenothera, was later exreached30-40% on Erigeronspecies(125). This difference were more (111) showedthat gravidparasitoids plainedwhen Shahjahan nectarincreasedtheir attractedto Erigeronflowerswherehigher-quality that Lygusbugson Erigeron the probability life span.This wouldincrease by a parasitoid. will be encountered cueshasbeenreviewed as searching useplantchemicals How parasitoids may act either directly as they by Vinson (141). These phytochemicals gut and volatilizefromthe plant,or aftertheypassthroughthe herbivore's are excreted[see (93) for examples]. Herbivore plantspecies[e.g. (11, attackrateson different different speciesexperience partsof the sameplant(150, 151). in (151)]or on different 60) andreviewed attack rates have been behind these differential Where the mechanisms individuals may escape it has usuallybeenfoundthat herbivore examined, attackby occupyingstationsthat are (a) seldom or never searchedby to enemies;or (c) located on a plant part that enemies;(b) inaccessible movement.Due to the smallsize and to enemysearching offersresistance of manyparasitic andpredaceous behavior insects,small highlystereotyped amountsof plant variationbetween species, within species, and within individuals can have a largeimpacton enemyforagingefficiency. patternsof many small enemiesoften restrict The specialized searching partsof a singleplantspecies. themto certainplantspeciesor to particular to parasiSincethe foliageof different plantspeciesis not equallyattractive run different risks of attack dependingon which plant toids, herbivores specializetheir search speciesthey occupy(e.g. 80, 82). Someparasitoids cones larvaefeedingin staminate budworm to specificplantorgans.Spruce the gallsufferhigherparasitism than those in vegetativebuds (35), and speis attackedby fewerparasitoid quercus-baccarum, maker,Neuroterus cies when its galls occur on oak catkinsthan when on leaves, the usual
HERBIVORE POSITION AND ENEMY SEARCH PATTERNS

50

et al PRICE

riskof location(2). On an evenfinerscale,positionon a leaf can influence path bug, Anthocoris confusus,follows a searching attack.The predatory along leaf marginsand midribs,and aphidsfeedingin these locationsare is Massalongia betulifolia, to attack(41). The gall-maker, morevulnerable when locatedon the leaf midrib(25%) less often attackedby parasitoids than when on the leaf lamina(44%) (3). havenot evolvedto searchin these whyenemies to determine It is difficult was made by Monteith(82) when he accessiblelocations.A contribution of two tachibetweenthe searching preferences founda positivecorrelation pine sawfliesfeeding on red-headed and theirsurvivorship nid parasitoids betweenplants and differences on differentplant species.Microclimatic (151). Enemiesmaybe adapted betweenplantpartsmayalsobe important theyarenot profitable. accessible locationssimplybecause to neglectcertain Finally,the host may recentlyhave expandedits resourceutilizationand the parasitoids may not yet have adaptedto this rangeextension. insectsareconcealed of herbivorous STRUCTURAL REFUGES A number to parasitoid penebarriers offering fromtheirenemiesby plantstructures in the size and complexity of theseplantpartsmakesfor tration.Variation of the herbivores usingthem. Suchstructural in the accessibility variation differences amongthe cultivarsof Brassicaoleraceaaffectthe parasitism Pierisrapae;attackratesare high on caulineleaves ratesof the herbivore, of open-growing varieties suchas brusselssprouts,and low on the heading escapesattackwhenit feeds wherethis caterpillar varieties suchas cabbage, in plantstructure are in the foldsof the leaves(87). Moresubtledifferences food plantsfor moths rateson different for different parasitism responsible which mine pine buds. When R. frustranavar. of the genus Rhyacionia, bushnellifeeds on pines with large buds and robuststems it suffersless parasitismthan when it feeds on small buds, probablybecausewhen it (55). A similar burrowsdeeplyit is beyondthe reachof small parasitoids of the European pine shoot mechanism explained why percentparasitism was much higher(25 times moth, R. buoliana,by Itoplectisconquisitor higherbetweenpure stands)in pupae collectedfrom Scots pine than in is too those collectedfrom red pine (1). The ovipositorof I conquisitor short to penetrate to the centerof many red pine buds. In addition,the of budsprotected by needleswashigherin redthanin Scotspine percentage (whereneedlesweresplayedat a greateranglefromthe stem).Burrowing is also importantin bark beetles,where beyondthe reachof parasitoids rates have been negativelycorrelatedwith bark thickness(5), parasitism and in insects attackinginflatedfruit (75). Porter (91) noted that when NorthAmerican speciesof applemaggotsfed in applesthey wereless often by Opiusmelleus than when they fed in smallernativefruits parasitized

INTERACTIONS AMONG THREETROPHICLEVELS

51

such as hawthorn becausethe ovipositorof the parasitoid could not reach hosts in the largerexotic fruit. Difficultyin reachingconcealedhosts might be overcomethroughthe in evolutionof longerovipositors. Althoughthis has undoubtedly occurred the coevolutionary race,we areunaware of any casesthat demonstrate this thatattackbruchid strategy. Parasitoids weevilsseemto haveovercome the problem through behavioral andphenological adaptation: Bruchids feeding on maturing legumeseedsare protected by thick,tough seed pods;parasitoids may thus be adapted to attackwhile the pod is still young,thin, and soft,or afterit hasfallento the groundandbegunto ferment (14, 15). Other enterthe pod throughthe hole boredby the ovipositing female parasitoids inface parasitoids bruchid(81). Similarproblems attackinggall-forming have sects,andsimilarphenological (144)andbehavioral (143)adaptations arisen.
INTERFERENCE WITH ENEMY SEARCH MOVEMENT Certainplants of small bearstructures andproduce secretions thatimpedethe movements predators and parasitoids in searchof prey. By increasing searchingtime suchplantstructures of enemies.Various weakenthe functional responses plant galls, for instance,produceglutinoussubstances(10, 32) that can immobilize smallinsectsand can presumably hinderparasitoids just when the gall is young and activelygrowing-i.e. when the gall-maker is most vulnerable to attack. The role of planttrichomes as a defenseagainstherbivores is well known [for reviewssee (74, 145)],but a dense mat of trichomescan also lower a normally effective enemyforaging efficiency. Encarsiaformosa, parasitoid of the greenhouse whitefly,is greatlyhinderedby the hairs producedby As a result,whitefly on thisplant(137, 153). cucumber. flourish populations In addition,honeydew on the hairs produced by the whiteflyaccumulates and then sticks to the parasitoid, amount forcingit to spenda significant of time grooming. The walkingspeedof E. formosa is threetimesgreater on glabrous thanon hairycucumber varieties (64);this fact makesbiological control feasibleon the former(39). Tobaccohairs producea viscous exudatethat impedesthe tiny parasitoids,Trichogramma minutum and Telonomus in reduced of Manduca sexta eggs sphingii, resulting parasitism of any physiological (69, 95). Independent functiontrichomesmay serve, theirlengthand densitymay evolveto a point wheredirectdeterrence to herbivory maybe heldat a submaximal levelby the counter-balancing force they exert on efficient foragingby the naturalenemiesof herbivores. PLANT TOXINS Many toxic chemicals,more or less unalteredfrom and parasitoids plants,may renderherbivorous insectstoxic to predators

52

PRICEet al

(102, 142)], (e.g. 59, 106).Chemicals maybe sequestered [e.g.cardenolides uptakeof comhemolymph composition maychangeowingto unavoidable pounds from the midgut (110), compoundsmay simply be in transit secretions throughthe gut (37), or plantchemicals maybe usedin defensive [see (38) and review(104)]. in the food consumed,which very likely resultsin important Variation in the chemicalcompositionof the herbivore,can affect its differences to predatorattack.Jays will consumethe grasshopper, Posusceptibility but not on Asclepiasspp. that conekilocerus bufonius,fed on dandelions tain cardenolides(105). The defensiveregurgitantof the grasshopper, Romalea microptera, effectivelydeterredants when its foodplant was or Salix nigra,but not whenit had eatenlettuce Eupatorium capillifolium butterflies, Danausplexip(37). Broweret al (17, 18)showedthatmonarch of palatability to jays when rearedon different pus, exhibiteda spectrum individuals(differenteither intra- or interspecifically); asclepiadaceous content. correlated palatability negativelywith cardenolide [see(47) for Toxinsderived fromplantscan also affectlarvalparasitoids them. Jones to detoxifyor sequester review],which selectfor adaptations (68) found that two parasitoids attackingcyanogenicZygaena spp. produced rhodanese,an enzyme that detoxifiescyanogeniccompounds.In a tachinidthatattacksthe monarch butterfly, addition,Zenilliaadamsonii, sequesterscardenolidesin its body (98), and a species of Microplitis by the (Braconidae) sequesters pyrrolizidine alkaloidsoriginallyproduced host plant (8). herbivore's Toxins derivedfrom plantsby herbivores can directlyaffectparasitoid Smith (118) found unusuallyhigh larval mortalityof two survivorship. of the California red scale,Aonidiella aurantii,when the scale parasitoids to whenit fed on citrushosts(100%compared fed on sagopalmcompared to 12%or less for the to less than 5% for one species,and 44% compared of Apanteles was determined. Larvalsurvivorship other);no toxic principle was drastically loweredwhenits host, the tobaccohornworm, congregatus to one withlow nicotinecontent[see wasfed a high-nicotine dietcompared of (134) and references therein].Since nicotineoccursin the hemolymph hornworms that this was the causal feedingon tobacco(110) it is probable & Duffy(23) demagent.A moredetailedexamplepresented by Campbell onstratedthe effectson the ichneumonid,Hyposoter exiguae, of adding tomatineto the artificialdiet of its host, Heliothiszea. Parasitoidlarval was slowed; development percentpupaleclosion,adultweight,andlongevity were reduced; and extractsof larvaecontainedtomatine. If toxinsarepresent in herbivore formidapotentially bodies,theypresent to successful attackby generalist andto colonization blebarriers parasitoids of parasitoid in evolutionary species time,as indicated by the largernumber

INTERACTIONS AMONG THREETROPHICLEVELS

53

on late-successional herbivores to early-successional compared speciesreportedabove.

EVOLUTIONARY CONSEQUENCES FOR HERBIVORES Plant influences

on enemyeffectiveness can have important for evolutionary consequences the herbivores involved.As we have documented above,these insectswill frequently experience lowervulnerability to enemieswhen feedingon one plant speciesor structure than another.With regardto birdsand cryptic prey,Brower that(exceptwherepreyspeciesdiversity (16) suggested is low) selectionwill limit the similaritybetweenprey species density-dependent and restrictthe diet breadth of any one speciesto those plantson whichit is best protected. As an interesting that live example,desertgrasshoppers on and mimic stems (stems exhibitlow interspecific variation)are more thanthosethatresemble polyphagous foliage(foliagevariesmorefromone speciesto another) the backgrounds (84).Ricklefs& O'Rourke (101)regard and the morphological that matchthem (the "escapespace") adaptations as a limitingresourcefor which speciescompete.The amountof escape space will partlydetermine herbivore speciesdiversityin a community. With some modifications, the same arguments apply when considering invertebrate enemiessuch as parasitoids that differfrombirdsin theiruse of more specificchemicaland physicalcues in host finding.Thus escape not only by mimeticappearances, but also by any spacecan be described of a widevarietyof nicheparameters to host finding andoviposiimportant tion behavior by enemies(e.g. positionon leaf, gall shape,odorsproduced for escapespace duringfeeding,foodplantspecies,etc) (158). Competition or "enemy-free selectionby space"(71) resultingfrom density-dependent such enemies is probablymore importantthan competitionfor food in the tremendous nichediversity generating foundin herbivorous insects,as illustrated by Askew's(2) studieson cynipidwasps. All plantpreferences by phytophagous insectsareprobably influenced to someextentby the effectiveness of enemiesassociated with the food items. For generalist restriction in dietbreadth or changesin the order herbivores, of foodplant canresultat leastas muchfromdifferential vulnerpreferences in foodplant as fromdifferences abilityto enemies quality(52). Smiley(116) suggeststhat predation and ants may explainwhy pressure by parasitoids Heliconius is specific melpomene to Passiflora oerstedii despiteits abilityto grow "aboutequallywell' on four other Passifloraspecies.(However,his datashowthatgrowthratesaresignificantly loweron at leastthreeof these thismightproduce species; an important difference overevolutionary time.) On the otherhand,host plantrangeextensions maybe reinforced by enemy ineffectiveness on a novel plant species. An extremecase concernstwo of the genus Tildeniaon hosts growingtogether closelyrelatedleafminers

54

PRICE et al

(P. Gross, unpublished). In one year 42% of the leafminers on Solanum carolinense were attacked by 25 species of parasitoids, while on adjacent Physalis heterophyllaonly one parasitoid species attacked 5% of the other leafminers. Such events have undoubtedly occurred when agriculturalcrops are imported. Native herbivores can expand their host ranges before enemies evolve the correspondinghost-finding abilities, resulting in creation of new pest species. For example, cutworms (e.g. Perosagrotis spp.) prefer cultivated grains over the native grass, Agropyron smithii, but their parasite, the tachinid Goniacapitata, still lays its microtype eggs on A. smithii (126). For specialized herbivores in the early stages of a speciation event via a host plant shift, serendipitousacquisition of enemy-free space may increase the viability of the founders (21, 91). On the other hand, host shift "attempts" may be throttled by enemy faunas already associated with the new plants. This is suggested by a review (54) indicating that enemies have interferred with as many as half the reported cases of herbivorous insects introduced for weed control.

PROPERTIES OF PLANT POPULATIONS

The properties of plant populations that influence the third trophic level have been poorly studied. Because much more research is needed in this area, it will not reward a thorough review at this time. Plant density and patch size influence herbivores and enemies alike [(85, 86); but cf (103)]. The most obvious effect of increased plant density and patch size is to increase resources for herbivoresto which many species respond positively, with subsequent density-dependentresponses by their enemies. Conversely, plants may occur at such a low density that specialized enemies find few specific herbivores, resulting in little or no parasitism (66). Plant density also influences microclimate, nutritional quality of foliage, and concentration of nectar and honeydew-all factors that effect members of higher trophic levels. One important effect of plant density is its impact on the diversity of associated plants that influence the upper trophic levels profoundly. This is discussed in the next section.

PROPERTIES OF PLANT COMMUNITIES

Emergent properties, which may well influence herbivore-enemy interactions, appear when plant species co-occur. The degree of association of a host plant with other plants [see (4) for the effect on the plant-herbivore relationship] affects three-trophic-level interactions in two major ways: (a) The herbivore-enemyinteractions on one plant species can be influenced

INTERACTIONS AMONG THREETROPHICLEVELS

55

by the presence of associated plant species; (b) the herbivore-enemyinteractions on one plant species can be influenced by the presence of herbivores on associated plant species. Plants with which the host plant is associated can be nectar and pollen sources for adult parasitoids and predators (e.g. 57, 113, 138, 154). Such nutrient sources frequently result in increased longevity and fecundity of enemies (72, 128, 130, 131), which may strengthen their functional and numerical responses [as suggested in (57)]. Quantitative studies are mostly on parasitoids(72, 73, 130, 131, 136), but effects on predatorshave also been noted (128). Associated plants supporting honeydew-producing herbivores are another important nutrient source for many enemies (26, 30, 61, 157); longevity and fecundity of the enemies may be increased (58, 157). The quantitativeeffects vary according to plant species and herbivoreproducing scutellathe honeydew (157). Some parasitoid species, such as Coccophagus ris, elicit honeydew secretion from homopterans as in ant trophobiosis (26). Gardener (50) found that Tiphiavernalis adults use honeydew produced by aphids and that Japanese beetle larvae were more heavily parasitized near plants infested with aphids than in areas free of aphids. Associated plants may provide odors or other conditions attractive to enemies. Read et al (97) found that aphids on sugar beets adjacentto collard plants were more frequentlyparasitizedby Diaeretiella rapae than those on sugar beets not near collards. The parasitoid normally attacks hosts on crucifersand apparentlysearched all plants in the vicinity of a collard plant. Smith (117) showed that some syrphid and anthocorid predators of aphids colonized brussels sprout plants among weeds more readily than plants on bare ground. This behavioris probablya function of the predators'tendency to move from plant to plant and the ameliorated conditions provided by the weedy vegetation. Alternatively, associated plants may interfere with the attractiveness of a plant to herbivores and their enemies, either by masking attractive odors from the host plant or by a direct repellent effect of the volatile chemicals of the associated plant. Tahvanainen & Root (132) found that nonhost plants such as tomato and common ragweed, when associated with the crucifer Brassicaoleracea, interferewith the searching and feeding behavior of the specialist herbivore Phyllotreta cruciferae on Brassica. They concluded that "the environmental capacity of diverse natural communities is lower than that of natural or man-made monocultures. The 'associational resistance' resulting from the higher taxonomic and microclimatic complexity of natural vegetation tends to reduce outbreaks of herbivores in diverse communities." It is likely that many specialist parasitoids are influenced by masking or repellent chemicals produced by associated plants.

56

PRICE et al

An example of this associational resistance to parasitoids centers on studies of host location by the tachinid ffies, Bessa harveyi and Drino bohemica (83). Many trees and forest understory plants apparently interfere with the host location abilities of the parasitoidsby masking the odor of the true host plant (Larix laricina). Parasitism of the host sawfly, Pristiphoraerichsonii, when Larix was associated with other plants, was lower (10-13%) than in pure stands of Larix (up to 86%). Although the herbivore responded to vegetation in a way similar to that of the parasitoids (avoiding Larix branches in contact with other plant species), those that did feed in these less apparent sites contributed disproportionately to the next generation because food was abundantand parasitism was low. Indeed, such unapparent sites for herbivoresmay act as important refugia from enemies. Similar patterns of parasitism have been observed on herbivores of herbaceous plants (100, 112). Associated plants may also support herbivoresthat act as alternate hosts, as indicated in studies of biological control and insect population dynamics (e.g. 20, 46, 63, 90, 113, 123). On associated plants enemies may remain at high population levels when the primary herbivorous host is at low population levels or when enemies are not well-synchronized with herbivore phenology. A well-documented example is that of the egg parasitoid, Anagrus epos (34). Wild blackberriessupport the herbivore, Dikrella cruentata (a leafhooper), which is attacked by Anagrus. Wild grape, growing in the same habitat, is attacked by another leafhopper, Erythroneuraelegantula. However, the leaf flush of wild grapes occurs just when the Anagrus adults are emerging after the first generation attack on Dikrella, so the parasitoid is present to colonize grape and attack the Erythroneura eggs as soon as they are laid. Atsatt & O'Dowd (4) suggest that selection may have favored those grape plants with a delayed leaf production that was synchronous with the availability of parasitoids.

THE PLANT'S PERSPECTIVE: A PARADOX

The causal relationships dealt with to this point have been those directed from the first trophic level towards the second and third. Superficially, at least, it would seem that the consequences of these for the plant would necessarilybe in a direction opposite to their consequences for the herbivore (excluding, of course, herbivore mutualists such as pollinators). However, more careful examination reveals that mechanisms that produce negative effects on the herbivore do not automatically lead to positive effects on the plant. Such paradoxicalsituations are best understood from the perspective of individual plant fitness.

INTERACTIONS AMONG THREETROPHICLEVELS

57

One such paradox arises upon consideration of the consequences of digestibility-reducers for plant-herbivore interactions. In general, insects rearedon diets of low digestibility(for example, high cellulose content) have a strong tendency to compensate through increased consumption (6). Similar compensation can be expected in response to plant digestibility-reducers (43). This means that the per capita damage by those insects that complete maturation on digestibility-reducing plants is actually greater because of those compounds. Therefore, from a plant-fitness perspective, the fact that digestibility-reducerscan increase enemy efficacy may be more important than originally conceived. Instead of being a supplement to the positive selective value of a digestibility-reducingtrait, enhanced enemy efficacymay be essential to making the value positive in the first place. Bouton et al (12) tested the importance of relative enemy efficacy in the selection for digestibility-reducingtraits. Identical numbers of larvae of the Mexican bean beetle, Epilachna varivestis, were introduced to cages containing plants of one of two soybean lines. Laboratory studies (70) had previously revealed differencesin the performance of the beetles (including growth rate and the ability to convert plant biomass to beetle biomass) on the two lines; these differencesclosely paralleled those between winter moth larvae on diets with and without tannin (42). The cause of these differences is unknown. Survivorshipto pupation on the variety PI 80837 (the "hightannin analog") was only 0.88 of that on the Harosoy Normal variety. However, since each larva on PI 80837 destroyed 1.26 times as much leaf area, this line actually sustained an estimated 1.11 times as much damage (1.26 X .88). The situation was reversed in cages where the predatory pentatomid, Podisus maculiventris, was introduced. Survivorship on PI 80837 was now only 0.23 of that on Harosoy Normal (predators were able to consume more of the slower growing larvae), so the overall damage was only 0.29 (i.e. 1.26 X .23) of that on Harosoy Normal. So the level of predation can determine the relative amounts of herbivore damage on plants with and without sublethal traits. Plant traits that facilitate parasitoid attack constitute a second category of mechanisms that negatively affect herbivores without automatically increasing plant fitness. Such traits can be expected to benefit the plant population as a whole by contributing to lower herbivore densities. But in order to become fixed in the population they must somehow enhance individual plant fitness, presumably by reducing localized herbivore damage. This is by no means a universal feature in parasitoid attack. The effect of parasitoid activity on localized herbivory strongly depends on the particular herbivore-parasitoidcombination. Parasitism does not always affect host feeding; many internal parasitoids that attack larvae are relatively inert within the host until after it has completed feeding (e.g. 94).

58

PRICE et al

Among parasitoidsthat do affect host feeding there is a spectrum of effects. Some hosts may quickly terminate feeding as a result of paralyzationby the ovipositing female or through rapid consumption by the immature parasitoid [e.g. some Banchini (Ichneumonidae) (92)]. In other cases, host feeding may continue in spite of internal parasitoid activity. This can result in reduced consumption (13, 129), but it can also result in an increase (65, 114); and in the case of Pieris rapae, consumption may either increase or decrease, depending upon whether they are attacked by gregarious or solitary parasitoids, respectively (96). The examples involving unchanged or increased herbivore feeding suggest that in some cases plants may only be able to benefit from higher parasitism by mechanisms involving parasitoidhost population dynamics. Whether such dynamic processes occur on a scale sufficientlylocalized to benefit individual plants requirescareful investigation. Thus plant traits have important consequences for herbivores and their enemies, but we need to investigate more closely the consequences for plant fitness. Plant fitness should be central to the study of the adaptive function of purported plant defenses; emphasis has traditionally been placed on the effects of defenses on herbivores.

CONCLUSIONS
We have argued for a holistic approach to plant-herbivore interactions. Considerationof the third trophic level is indispensableto an understanding of any part of the system. We cannot understand the plant-herbivore interaction without understandingthe role of enemies. We cannot understand predator-preyrelationships without understanding the role of plants. Enemies should be considered as mutualists with plants and part of plant defense. They can be manipulatedby natural selection on plants with compromises made between intrinsic defense and traits that foster effective enemies. Thus good intrinsic plant defense frequently results in negative impact on enemies, and a lowering of these defenses may benefit enemies and plant fitness. Plant fitness must be recognized as a central theme. The costs and benefits must be measured and ultimately incorporated into a quantitative theory of plant-herbivore-enemyinteractions. Emphasis on the third trophic level extends the theory of plant-herbivore interactions and broadensits basis. Plant fitness is fundamental in considering plant-herbivore interactions. Acknowledgment that the plant is the basic level at which natural selection operates in the three-trophic-level system avoids paradoxicalargumentssuch as those discussed in this review. Theories on population dynamics, biological control, and the actual num-

INTERACTIONS AMONG THREETROPHICLEVELS

59

bers of parasitoid species on early- versus late-successional herbivores are all consistent with Feeny's prediction that enemies are favoredby persistent plants with quantitative defenses. Further studies should proceed with a view to furtheringtheory on plant-herbivore-enemyinteractions ratherthan on a two-trophic-level system. Implications of the three-trophic-level theory for agriculture are extensive. Many crop plants cultivated in monocultures are actually derived from early-successionalspecies:Unapparent plants have been made apparent(44, 45). Manipulation of this vulnerable system to reduce herbivore impact must be subtle and based on detailed knowledge. Increased diversity within a field, either genotypic or plant species diversity, would decrease plant apparency (45), but associated plants may mask cues used by enemies for searching and reduce their efficiency. Breeding programs to increase plant resistance to herbivores may well select for lines that confer increased resistance of herbivores to their enemies. Manipulations for the biological control of weeds should attempt to reduce efficacy of enemies to enable a strong numerical response of herbivoresto their superabundantplant food. Again cost-benefit analysis will help to optimize management practices; such optimization cannot be achieved without considering the enemies of herbivores. Many areas of research need more concentrated effort. Plants probably always influence the third trophic level, but we need to understand better what aids and hinders enemies. The role of enemy-free space needs more exploration. One common effect of plants will be through microorganisms in herbivores, particularlythe gut microorganisms (e.g. 67) and mutualists elsewhere in the body. Digestibility of food, vulnerability to pathogens, and plant specificity may all be mediated by associated microorganisms. Isogenic lines of plant species are available to help unravel the impact of one plant trait at a time on herbivores and enemies, providing a potent tool for analyzing interactions. Tropical systems must be integrated into the theory developed largely on knowledge from temperate latitudes. The challenges are great for the development of a quantitative theory on plant-herbivoreenemy interactions, but the potential rewards are also considerable.
ACKNOWLEDGMENTS

We thank Clive Jones for reviewing the manuscript and for illustrating the importance of plant toxins for microorganisms in herbivores through his research on silk moths. Bruce Campbell kindly provided some valuable references and reprints, and he and Murray Isman, David Seigler, Paul Feeny, Daniel Hare, Bradleigh Vinson, and Ronald Weseloh read the
review.

60

PRICE et al

Literature Cited

1. Arthur, A. P. 1962. Influence of host tree on abundanceof Itoplectis conquisitor (Say) (Hymenoptera: Icha polyphagous parasite of neumonidae), the European pine shoot moth, Rhya337-47 2. Askew,R. R. 1961.On the biologyof theinhabitants of oakgallsof Cynipidae in Britain.Trans.Soc. (Hymenoptera) 3. Askew, R. R., Ruse, J. M. 1974.The biologyof some Cecidomyiidae (Diptera)gallingthe leavesof birch(Betula) with special referenceto their chalTrans. cidoid(Hymenoptera) parasites.
Brit. Entomol. 14:237-68 cionia buoliana (Schiff.) (Lepidoptera: Olethreutidae). Can. Entomol. 94:

14. Bridwell, J. C. 1918.Notes on the Bruin the Hawaichidaeandtheirparasites ian Islands. Proc. HawaiianEntomol. 15. Bridwell, J. C. 1920.Notes on the Bruand theirparasites chidae(Coleoptera) in the HawaiianIslands. 3rd paper. Entomol. Soc. 4:403-9 Proc.Hawaiian 16. Brower, L. P. 1958.Birdpredation and in closely related foodplantspecificity procryptic insects.Am. Nat. 92:183-87 17. Brower,L. P., Brower,J. V. Z., Corvino,J. M. 1967.Plantpoisonsin a terrestrialfood chain. Proc. Natl. Acad.
Sci USA 57:893-98 Soc. 3:465-505

18. Brower,L.P., Ryerson,W. N., Coppinger,L. L., Glazier,S. C., 1968.EcoR. Entomol. Soc. London 126:129-67 logical chemistryand the palatability 4. Atsatt, P. R., O'Dowd, D. J. 1976. spectrum. Science 161:1349-51 Plantdefense guilds.Science193:24-29 19. Bucher,G. E. 1960.Potential bacterial D. L. 1973.Hyme5. Ball,J. C., Dahilsten, of insectsand their characpathogens nopterous parasites of Ips paraconfusus teristics.J. InsectPathol.2:172-95 (Coleoptera:Scolytidae) larvae and 20. Ilurleigh, J. G., Young,J. H., Morrison, to mortality. theircontribution I. InfluR. D., 1973. Strip-cropping's effecton ence of host tree and treediameter on beneficial insectsandspiders associated parasitization. Can. Entomol. 105: Environ. Enwith cottonin Oklahoma. 1453-64 tomol.2:281-85 6. Barton Browne,L. 1975. Regulatory 21. Bush, speciation G. L. 1975.Sympatric in insectfeeding. mechanisms Rec.Adv. in phytophagous parasiticinsects. In Insect Physiol. 11:1-116 Inof Parasitic Evolutionary Strategies 7. Beddington,J. R., Hammond,P. S. sects and Mites, ed. P. W. Price, pp. 1977.Onthedynamics of host-parasite187-206.NY: Plenum hyperparasite interactions.J Anim. 22. Cagampang,G. B., Pathak, M. D., Ecol. 46:811-21 B. 0. 1974.Metabolic changes Juliano, 8. Benn, M., Degrave, J., Gnanasunin the rice plant duringinfestation by deram, C., Hutchins,R. 1979. HostNilaparvata luthe brownplanthopper, plantpyrrolizidine alkaloids in Nyctemgens Stal (Hemiptera:Delphacidae). era annulata Boisduval: Their persisAppl.Entomol.Zool. 9:174-84 thelife-cycle tencethrough andtransfer 23. Campbell,B. C., Duffy, S. S. 1979. to a parasite. 35:731-32 Experientia andparasitic Tomatine wasps: potential nectar9. Bentley,B. L. 1977.Extrafloral of plantantibiosis with incompatibility ies andprotection by pugnacious bodybiologicalcontrol.Science205:700-2 guards. Ann. Rev. Ecol. Syst. 8:407-27 I. M. 1962.Reproductive ca10. Bequaert,J. 1924. Galls that secrete 24. Campbell, Led. pacityin the genusChoristoneura A contribution to the probhoneydew. I. QuantitaTortricidae). (Lepidoptera: lem as to whethergalls are altruistic andgenesas controllers tiveinheritance adaptions. Bull. BrooklynEntomol. Soc. of rates.Can.J. Genet. Cytol.4:272-88 19:101-24 25. Caswell, H., Reed,F. C. 1975.Indigest11. Bombosch, S. 1966.Occurrence of eneibilityof C4 bundlesheathcells by the mieson different weedswithaphids.In grasshopper, Melanoplus confusus. Ecology ofAphidophagousInsects, ed. I. Ann.Entomol.Soc.Am. 68:686-88 Hodek,pp. 177-79.The Hague:Junk S. M. 1937. Studieson the 26. Cendania, 12. Bouton,C. E., Price,P. W., Kogan,M., (Hymenopbiology of Coccophagus 1981.Chemical in plantsandits defense tera), a genus parasiticon nondiaspirelevance to thethirdtrophic level.Am. dine Coccidae. Univ. Calif Publ. EnNat. In press tomol.6:337-400 F. D., King,E. G. 1978.Effects 13. Brewer, of parasitism by a tachinid, Lixophaga 27. Chan, B. G., Waiss,A. C., Lukefahr, M. 1978.Condensed tannin,an antibidiatraeae,on growth and food conotic chemical from Gossypiumhirsumption of sugarcaneborer larvae. Ann. Entomol. Soc. Am. 71:19-22 sutum.J. InsectPhysiol.24:113-18

INTERACTIONS AMONG THREE TROPHIC LEVELS

28. Cheng, C. H., Pathak, M. D. 1972. Resistance to Nephotettix virescensin rice varieties. J. Econ. Entomol. 65:1148-53 29. Cole, L. C. 1954. The population consequences of life history phenomena. Q. Rev. Biol. 29:103-37 30. Cole, L. R. 1967. A study of the lifecycles and hosts of some Ichneumonidae attacking pupae of the green oak-leaf roller moth Tortrix viridana (L.) (Lepidoptera: Tortricidae) in England. Trans. R. Entomol. Soc. London 119:267-81 31. Dadd, R. H. 1973. Insect nutrition:current developments and metabolic implications. Ann. Rev. Entomol. 18: 381-420 32. Darlington, A. 1975. The Pocket Encyclopaedia of Plant Galls in Colour. Poole, Dorset: Blandford. 191 pp. Rev. ed. 33. Dodd, G. D. 1973. Integratedcontrol of the cabbage aphid, Brevicoryne brassicae (L.) PhD thesis. Univ. Reading 34. Doutt, R. L., Nakata, J. 1973. The Rubus leafhopper and its egg parasitoid: an endemic biotic system useful in grape-pest management. Environ. Entomol. 2:381-86 35. Dowden, P. B., Carolin, V. M., Dirks, C. 0. 1950. Natural control factors affecting the spruce budworm in the Adirondacks during 1946-1948. J. Econ. Entomol. 43:774-83 36. Edmunds, G. F., Alstad, D. N. 1978. Coevolution in insect herbivoresand conifers. Science 199:941-45 37. Eisner, T. 1970. Chemical defense against predation in arthropods. In Chemical Ecology, ed. E. Sondheimer, J. B. Simeone, pp. 157-217. NY: Academic 38. Eisner, T., Johnessee, J. S., Carrel, J., Hendry, L. B., Meinwald, J. 1974. Defensive use by an insect of a plant resin. Science 184:996-99 39. Ekbom, B. S. 1977. Development of a biological control program for greenhouse whiteffies (Trialurodes vaporarium Westwood) using its parasite Encarsia formosa (Gahan) in Sweden. Z. Angew. Entomol. 84:145-54 40. Erickson, J. M., Feeny, P. 1974. Sinigrin: A chemical barrier to the black swallowtail butterfly, Papiliopolyxenes. Ecology 55:103-11 41. Evans, H. F. 1976. The searching behaviour of Anthocorisconfusus (Reuter) in relation to prey density and plant surface topography. Ecol. Entomol. 1:163-69

61

42. Feeny, P. P. 1968. Effect of oak leaf tannins on larval growth of the winter moth Operophterabrumata. J Insect Physiol. 14:805-17 43. Feeny, P. P. 1970. Seasonal changes in oak leaf tannins and nutrients as a cause of spring feeding by winter moth caterpillars. Ecology 51:565-81 44. Feeny, P. P. 1975. Biochemical coevolution between plants and their insect herbivores. In Coevolution of Animals and Plants, ed. L. E. Gilbert, P. H. Raven, pp. 3-19. Austin: Univ. Texas Press 45. Feeny, P. P. 1976. Plant apparency and chemical defense. Rec. Adv. Phytochem. 10:1-40 46. Flaherty, D. 1969. Ecosystem trophic complexity and densities of the Willamette mite, Eotetranychuswillamettei (Acarnia: Tetranychidae). Ewing Ecology 50:911-16 47. Flanders, S. E. 1942. Abortive developmerit in parasitic Hymenoptera, induced by the food-plant of the insect host. J. Econ. Entomol. 35:834-35 48. Fraenkel, G. 1953. The nutritional value of green plants for insects. Proc. 9th Int. Congr. Entomol. 2:90-100 49. Futuyma, D. J. 1976. Food plant specialization and environmental predictability in Lepidoptera. Am. Nat. 110:285-92 50. Gardner, T. R. 1938. Influence of feeding habits of Tiphia vernalis on parasitization of the the Japanese beetle. A Econ. Entomol. 31:204-7 51. Gilbert, L. E. 1975. Ecological consequences of a coevolved mutualism between butterfliesand plants. In Coevolution of Animals and Plants, ed. L. E. Gilbert, P. H. Raven, pp. 21040. Austin: Univ. Texas Press 52. Gilbert, L. E. 1979. Development of theory in the analysis of insect-plant interactions. In Analysis of Ecological Systems, ed. D. J. Horn, G. R. Stairs, R. D. Mitchell, pp. 117-54. Columbus, Ohio: Ohio State Univ. Press 53. Gilbert, L. E., Smiley, J. T. 1978. Determinants of local diversity in phytophagous insects: host specialists in tropical environments. Symp. R. Entomol. Soc. London 9:89-104 54. Goeden, R. D., Louda, S. M. 1976. Biotic interference with insects imported for weed control. Ann. Rev. Entomol. 21:325-42 55. Graham, S. A., Baumhofer, L. G. 1927. The pine tip moth in the Nebraska National Forest. J. Agric. Res. 35:323-33

62

PRICE et al
eggs of the tobacco hornworm on various host plants in Kentucky. Environ. Entomol. 2:939-45 Kogan, M. 1972. Intake and utilization of natural diets by the Mexican bean beetle, Epilachna varivestis-a multivariate analysis. In Insect and Mite Nutrition, ed. J. G. Rodriquez, pp. 10726. Amsterdam: North Holland Lawton, J. H. 1978. Host-plant influences on insect diversity: the effects of space and time. In Diversity of Insect Faunas, ed. L. A. Mound, N. Waloff, pp. 105-25 Oxford: Blackwell Leius, K. 1963. Effects of pollens on fecundity and longevity of adult Scambus buolianae (Htg.) (Hymenoptera: Ichneumonidae). Can. Entomol. 95: 202-7 Leius, K. 1967. Influence of wild flowers on parasitism of tent caterpillar and codling moth. Can. Entomol. 99: 444 46 Levin, D. A. 1973. The role of trichomes in plant defense. Q. Rev. Biol. 48:3-15 Lewis, W. J., Brazzel, J. R., Vinson, S. B. 1967. Heliothis subflexa, a host for Cardiochelis nigriceps. J. Econ. Entomol. 60:615-16 May, R. M. 1971. Stability in model ecosystems. Proc. Ecol. Soc. Aust. 6:18-56 May, R. M. 1972. Will a large complex system be stable? Nature 238:413-14 Maynard Smith, J. 1974. Models in Ecology. London: Cambridge Univ. Press. 146 pp. McNeil, S., Southwood, T. R. E. 1978. The role of nitrogen in the development of insect/plant relationships. In Biochemical Aspects of Plant and Animal Coevolution,ed. J. B. Harborne,pp. 7798. London: Academic Miller, C. A. 1959. The interaction of the spruce budworm, Choristoneura fumiferana (Clem.), and the parasite Apanteles fumiferanae Vier. Can. Entomol. 91:457-77 Mitchell, R. 1977. Bruchid beetles and seed packaging by palo verde. Ecology 58:644 51 Monteith, L. G. 1955. Host preferences of Drino bohemica Mesn. (Diptera: Tachinidae), with particular reference to olfactory responses. Can. Entomol. 87:509-30 Monteith, L. G. 1960. Influence of plants other than the food plants of their host on host-finding by tachinid parasites. Can. Entomol. 92:641-52

56. Griffiths,K. J. 1969. The importance of coincidence in the functional and numerical responsesof two parasitesof the European pine sawfly, Neodiprion sertifer. Can. Entomol. 101:673-713 57. Gybrfi, J. 1951. Die Schlupfwespenund der Unterwuchs des Waldes. Z. Angew. Entomol. 33:24-47 58. Hagen, K. S. 1950. Fecundity of Chrysopacalifornica as affectedby synthetic foods. J. Econ. Entomol. 43:101-4 59. Harborne, J. B. 1977. Introduction to Ecological Biochemistry. London: Academic. 243 pp. 60. Haynes, D. L., Butcher, J. W. 1962. Studies on host preferenceand its influence on EuropeanPine Shoot Moth success and development. Can. Entomol. 94:690-706 61. Herrebout, W. M. 1967. Habitat selection in Eucarcelia rutilla Vill. (Diptera: Tachinidae). I. Observationson the occurrence during the season. Z Angew. Entomol. 60:219-29 62. House, H. L. 1969. Effects of different proportions of nutrients on insects. Entomol. Exp. Appl. 12:651-69 63. Hsiao, T. H., Holdaway, F. G. 1966. Seasonal history and host synchronization of Lydella grisescens (Diptera: Tachinidae) in Minnesota. Ann. Entomol. Soc. Am. 59:125-33 64. Hulspas-Jordan,_P. M., van Lenteren, J. C. 1978. The relationshipbetween hostplant leaf structure and parasitization efficiencyof the parasiticwasp Encarsia Gahan (Hymenoptera: formosa Aphelinidae). Med. Fac. Landbouww. Rijksuniv. Gent 43:431-40 65. Hunter, K. W., Stoner, A. 1975. Copidosoma truncatellum: Effect of parasitization on food consumption of larval Trichoplusia ni. Environ. Entomol. 4:381-82 66. Janzen, D. H. 1975. Interactions of seeds and their insect predators/ parasitoids in a tropical deciduous forest. In EvolutionaryStrategies of Parasitic Insects and Mites, ed P. W. Price, pp. 154-86. NY: Plenum 67. Jones, C. G. 1980. Bald cypress allelochemics and the inhibition of silkworm enteric microorganisms:Some ecological considerations. J. Chem. Ecol. In press 68. Jones, D. A. 1966. On the polymorphism of cyanogenesis in Lotus corniculatus. I. Selection by animals. Can. J. Genet. Cytol. 8:556-67 69. Katanyukul, W., Thurston, R. 1973. Seasonal parasitism and predation of

70.

71.

72.

73.

74. 75.

76. 77. 78. 79.

80.

81. 82.

83.

INTERACTIONS AMONG THREE TROPHIC LEVELS

84. Otte, D., Joern, A. 1977. On feeding patterns in desert grasshoppersand the evolution of specialized diets. Proc.

63

Acad. Nat. Sci Philadelphia 128:

89-126 85. Pimentel, D. 1961. The influence of plant spatial patterns on insect populations. Ann. Entomol. Soc. Am. 54:61-69 86. Pimentel, D. 1961. Species diversity and insect population outbreaks. Ann. En-

in the monarch butterfly. Science 161:861-66 D. F., Cates,R. G. 1976.To99. Rhoades, warda general theoryof plantantiherbivore chemistry. Rec. Adv. Phytochem.

tomol.Soc. Am. 54:76-86

87. Pimentel, D. 1961. An evaluation of insect resistance in broccoli, Brussels sprouts, cabbage, collards and kale. J.

Econ.Entomol.54:156-58
88. Pimm, S. L., Lawton, J. H. 1977. Number of trophic levels in ecological communities. Nature 268:329-31 89. Pimm, S. L., Lawton, J. H. 1978. On feeding on more than on trophic level. Nature 275:542-44 90. Plakidas, J. D. 1978. Epiblema scudderiana (Clemens) (Lepidoptera: Olethreutidae), a winter host reservoir for parasitic insects in southwestern Pennsylvania. J. N. Y Entomol. Soc. 86:220-23 91. Porter, B. A. 1928. The apple maggot.

10:168-213 100. Richards,0. W. 1940.The biologyof the smallwhitebutterfly (Pierisrapae), withspecial reference to thefactors controllingits abundance. J. Anim.Ecol. 9:243-88 101. Ricklefs, R. E., O'Rourke, K. 1975.Aspect diversityin moths: a temperatetropical comparison. Evolution 29:313-24 102. Roeske,C. N., Seiber, J. N., Brower, L. P., Moffitt, C. M. 1976.Milkweed cardenolidesand their comparative probutterflies cessingby monarch (Danaus
plexippus L.). Rec. Adv. Phytochem.

10:93-167 103. Root, R. B. 1973. Organization of a plant-arthropod associationin simple and diversehabitats: the faunaof collards (Brassica oleracea). Ecol. Monogr.

43:95-124 104. Roth, L. M., Eisner, T. 1962.Chemical Ann. Rev. Endefensesof arthropods. M. 1966.Experiments 105. Rothschild, with captive predatorsand the poisonous C 31:32 106. Rothschild,M. 1973.Secondary plant and warning in colouration substances ed. H. F. van Emden,pp. 59-83. Oxford:Blackwell 107. Ryan,C. A. 1973.Proteolytic enzymes in plants.Ann.Rev. andtheirinhibitors 108. Scriber, J. M. 1977.Limiting effectsof low leaf-water contenton the nitrogen utilization,energy budget, and larval
growth of Hyalophora cecropia (LepiPlant Physiol. 24:173-96 insects. In Insect/Plant Relationships, grasshopper Poekilocerus bufonius (Klug.). Proc. R. Entomol. Soc. London tomol. 7:107-36

U S. Dep.Agric.Tech.Bull. 66:1-48

92. Price, P. W. 1975. Reproductive strategies of parasitoids. In Evolutionary

Strategies of Parasitic Insects andMites,

ed. P. W. Price, pp. 87-111. NY: Plenum 93. Price, P. W. 1980. Semiochemicals in evolutionary time. In Semiochemicals:

TheirRole in Pest Control, ed. D. A.

94.

95.

96. 97.

98.

Nordlund, R. L. Jones, W. J. Lewis. NY: Wiley. In press Price, P. W., Tripp, H. A. 1972. Activity patterns of parasitoidson the Swaine jack pine sawfly, Neodiprion swainei (Hymenoptera: Diprionidae), and parasitoid impact on the host. Can. Entomol. 104:1003-16 Rabb, R. L., Bradley, J. R. 1968. The influence of host plants on parasitism of eggs of the tobacco hornworm. J. Econ. Entomol. 61:1249-52 Rahman, M. 1970. Effect of parasitism on food consumption of Pieris rapae larvae. J. Econ. Entomol. 63:820-21 Read, D. P., Feeny, P. P., Root, R. B. 1970. Habitat selection by the aphid parasite Diaeretiella rapae (Hymenoptera: Branconidae) and hyperparasite Charips brassicae (Hymenoptera: Cynipid&e). Can. Entomol. 102: 1567-78 Reichstein, T., von Euw, J., Parsons, J. A., Rothschild, M. 1968. Heart poisons

doptera: Saturniidae). Oecologia 28:269-87 109. Scriber, J. M., Feeny,P. 1979.Growth in relation of herbivorous to caterpillars andto the growth feeding specialization form of their food plants. Ecology 60:829-50 110. Self,L. S., Guthrie,F. E., Hodgson,E. 1964. Adaptation of tobacco hornwormsto the ingestionof nicotine.J. 111. Shahjahan, M. 1974. Erigeronflowers as a foodandattractive odorsourcefor of the tarmshedplant bug. parasitoid
Environ. Entomol. 3:69-72 Peristenus pseudopallipes, a braconid Insect Physiol. 10:907-14

64

PRICE et al
in insect disease. Proc. 10th Int. Congr. Entomol. 4:725-30

112. Shahjahan, M., Streams,F. A. 1973. Planteffectson host-finding by Leiophron

Braconidae), a parasitoidof the tarnished plant bug. Environ. Entomol.

pseudopallipes

(Hymenoptera:

125. Streams, F. A., Shahjahan, M., H. G. 1968. Influenceof LeMasurier, plantson the parasitization of the tar-

nished plant bug by Leiophronpallipes. 2:921-25 J. Econ. Entomol. 61:996-99 113. Simmons, D. E., Chen, G. A., Leonard, E. H. 1923.Biological notes C. W. 1975. Influence of tree species 126. Strickland, on parasites of prairiecutworms.Can. densityand composition of parasitism Dep. Agr. Entomol. Br. Bull. 22:1-40 of the sprucebudworm, Choristoneura W. H., Langenheim, J. H. fumiferana (Clem.). Environ. Entomol. 127. Stubblebine, 1977. Effects of Hymenaea courbaril 4:832-36 leaf resin on the generalistherbivore 114. Slansky,F. Jr. 1978.Utilization of enSpodopteraexigua (beet armyworm). J. ergy and nitrogen by larvaeof the imChem. Ecol. 3:633-47 Pierisrapae, as ported cabbageworm, of food affected by parasitismby Apanteles 128. Sundby,R. A. 1967.Influence

Stephens (Neuroptera,Chrysopidae). 7:179-85 Entomophaga 12:475-79 115. Slansky, F. Jr., Feeny,P. 1977.StabiliW. E. 1978. G. A., Wallner, zationof the rateof nitrogen accumula- 129. Surgeoner, Foliage consumption by the variable tion by larvaeof the cabbage butterfly oak leaf caterpillar, Heterocampa on wildandcultivated foodplants. Ecol. manteo (Lepidoptera: Notodontidae), Monogr. 47:209-28 its use in defoliation predictions.Can 116. Smiley, J. 1978.Plantchemistry andthe Entomol. 110:241-44 evolutionof host specificity: New eviof flowers dence for Heliconius and Passiflora. 130. Syme,P. D. 1975.The effects on the longevityand fecundityof two Science 201:745-47 native parasitesof the Europeanpine 117. Smith,J. G. 1969.Someeffectsof crop shoot moth in Ontario.Environ.Enon populations of aphids background tomol. 4:337-46 and their naturalenemieson Brussels 131. Syme, P. D. 1977.Observations on the sprouts. Ann. Appl. Biol. 63:326-30 longevityand fecundityof Orgilus ob118. Smith,J. M. 1957.Effects of the food scurator (Hymenoptera: Braconidae) plantof California redscale,Aonidiella and the effectsof certainfoodson lonaurantii(Mask.), on reproduction of its gevity. Can. Entomol. 109:995-1000 hymenopterousparasites. Can. En- 132. Tahvanainen, J. O., Root, R. B. 1972. tomol. 89:219-30 The influence of vegetational diversity 119. Southwood, T. R. E. 1975.Thedynamon the population ecologyof a specialics of insect populations.In Insects, ized herbivore, Phyllotreta cruciferae Science, and Society, ed. D. Pimentel, (Coleoptera: Chrysomelidae). Oecologia pp. 151-99. NY: Academic 10:321-46 120. Southwood,T. R. E. 1977. The rele- 133. Theophrastus. into c. 300 B.C. Enquiry vanceof population dynamic theoryto Plants and Minor Workson Odoursand pest status. In Originsof Pest, Parasite, Weather Signs, Vol. 1. (Trans.Sir ArDisease and Weed Problems, ed. J. M. thurHart).Cambridge: Harvard Univ. Cherrett, G. R. Sagar.pp. 35-54. OxPress.475 pp. ford:Blackwell 134. Thurston, R., Fox, P. M. 1972.Inhibi121. Southwood, T. R. E., Comins,H. N. tion by nicotineof emergence of Apan1976.A synopticpopulation model.J. teles congregatus from its host, the toAnim. Ecol. 45:949-65

glomeratus.

Environ.

Entomol.

on the fecundity of Chrysopa carnea

122. Starks,K. J., Muniappan, R., EikenAm. 65:547-50 bary, R. D. 1972.Interaction between 135. Tilman, D. 1978. Cherries,ants and plantresistance and parasitism against tent caterpillars: Timingof nectarprogreenbug on barleyandsorghum. Ann. of ductionin relationto susceptibility Entomol. Soc. Am. 65:650-55 to ant predation.Ecology caterpillars 123. Stary, P. 1978. Seasonalrelationsbe59:686-92 tween lucerne,red clover, wheat and 136. Topham,M., Beardsley, J. W. Jr. 1975. barley agro-ecosystems through the Influence of nectarsourceplantson the aphids and parasitoids(Homoptera, weevilparasite, New Guineasugarcane Aphididae; Hymenoptera, Aphidiidae). Lixophaga sphenophori (Villeneuve).
Acta Entomol. Bohem. 75:296-311 Proc. Hawaiian Entomol. Soc.

bacco hornworm. Ann. Entomol. Soc.

124. Steinhaus, E. A. 1958.Stressas a factor

22:145-54

INTERACTIONS AMONG THREE TROPHIC LEVELS J. S., vanLenteren, 137. Vande Merendonk, of of mortality v. 1978. Determination
greenhouse whitefly Tiraleurodes va-

65

148. Weld,L. H. 1925.Anothergallthatsecretes honeydew. Bull. Brooklyn Entomol. Soc. 20:175-79

porarium (Westwood) (Homoptera: 149. Wellso,S. G. 1973.Cerealleaf beetle: development orientation, larvalfeeding, eggs,larvaeandpupaeon Aleyrodidae) cultiand survivalon four small-grain (Solaeggplant species: fourhost-plant Ann. Entomol. vars in the laboratory. num melongena L.), cucumber (Cucumis sativus L.), tomato (Lycopersicum esculentum L.) and paprika (Capsicum annuum L.). Med. Fac. Landbouww Rijksuniv. Gent 43:421-29

agilis. Environ. Entomol. 3:363-64 138. Van Emden,H. F. 1963.Observations of forest R. M. 1976.Behavior of 151. Weseloh, on theactivity on theeffectof flowers in Mon. insect parasitoids.In Perspectives Entomol. Hymenoptera. parasitic Forest Entomology, ed. J. F. Anderson, Mag. 98:265-70 H. K. Kaya,pp.99-110.NY:Academic 139. Van Emden,H. F. 1966.Plant insect to and pest control. World 152. White,T. C. R. 1974.A hypothesis relationships of loopercaterpillars, outbreaks explain Rev. Pest Control 5:115-23 of to populations with specialreference con140. Varley,G. C. 1959.Thebiological Selidosema suavis in a plantation of pests.J. R. Soc.Arts trolof agricultural Pinus radiata in New Zealand. 107:475-90 Oecologia 16:279-301 141. Vinson,S. B. 1976. Host selectionby J. C. 1976.The insect parasitoids. Ann. Rev. Entomol. 153. Woets,J., vanLenteren, parasite-host relationship between 21:109-33 Encarsia formosa (Hymenoptera: 142. Von Euw, J., Reichstein,T., RothAphelinidae) and Trialurodes vaacid-Iin schild,M. 1968.Aristolochic porarium (Homoptera:Aleyrodidae). the swallowtailbutterflyPachlioptera of the host planton VI. The influence Isaristolochiae (Fabr.)(Papilionidae). andits parasite whitefly the greenhouse rael J. Chem. 6:659-70 Encarsiaformosa. Proc. 3rd Conf: Biol. J. K. 1977.A new Tetrasti143. Wangberg, Control Glasshouses O.I.L.B./S.R.O.P. gall-formers tephritid chus parasitizing 76:125-37 on Chrysothamnusin Idaho.Pan-Pac. 154. Wolcott,G. N. 1942.Therequirements Entomol. 53:237-40 formorethanhosts.Science of parasites J. D., Cornell,H. V. 1979. 144. Washburn, 96:317-18 Chalcid parasitoidattack on a gall of wasp population (Acraspis hirta 155. Wyatt, I. J. 1965.The distribution Myzus persicae (Sulz.) on year-round on Quercus Cynipidae)) (Hymenoptera: I. Summerseason. chrysanthemums. prinus (Fagaceae). Can. Entomol. Ann. Appl. Biol. 56:439-59 111:391-400 of 145. Webster,J. A. 1975. Associationof 156. Wyatt, I. J. 1970. The distribution Myzus persicae (Sulz) on year-round an anplanthairsandinsectresistance: II. Winter season: Chrysanthemums. US. Dep.Agric. notatedbibliography. maby Aphidius the effectof parasitism Misc. Publ. 1297:1-18 tricariae Hal. Ann. Appl. Biol. 65:3141 146. Weld,L. H. 1922.Notes on American produc- 157. Zoebelein,G. 1956.Der Honigtauals Cynipidae of the family ga11ffies of Nahrungder Insekten.Z. Angew.Enwithdescriptions mg gallson acorns, tomol. 38:369416; 39:129-67 new species. Proc. US Natl. Mus. andniche H. 1975.Speciation 158. Zw6lfer, 61(19):1-32 in phytophagous insects. diversification fromoak 147. Weld,L. H. 1925.Honeydew galls. Am Bee J. 65:469 Verh.Dtsch. Zool. Ges. 67:394-401

mi150. Weseloh,R. M. 1974.Host-related crohabitatpreferencesof the gypsy moth larval parasitoid,Parasetigena

Soc. Am. 66:1201-8