t h e s c i e n c e o f

E N T O M O L O G Y

W i l l i a m S . R o m o s e r

O h i o Universi ty

J o h n C j . S t o f f o l a n o , J r .

Univers i ty of Massachusetts

Boston, Massachusetts Burr Ridge, lllinois Dubuque, Iowa

I've \\latched you now a full half-hour, Self-poised upon that yellow flotver; And, little butterfly! indeed I know not if you sleep or feed.

William Words\vonh, To a Buiierjl?

C H A P T E R lu

Humans have always been fascinz~ed by insects, and the majority of their observations occur on the substrate of plants. Plants, along \vith algae, f o m the basis of nearly all food chains on Earth because they rre the only organisms that capture energy from sunlight. In turn, herbivores (of which many a re insects) are plant consumerstthereby transfening energy to higher trophic levels (figure 10.1). Insects and plants have evolved together for millions of years, and their relationships include both antagonistic 2nd mutualistic interactions. Most early obsenlations were sim- ply accounts of plague insects (e.g., in the Bible); only during the last 75 years have insects on plants been more carefully studied. Why has the biology of insects on plants received such extensive recent attention? There are a num- ber of reasons:

1. Insect population fluctuations have a direct impact on human health and food supplies (see also Part 4, "Applied Entomology").

2. Insects are abundant, small, and easy to raise, obsene, and manipulate in scientific experiments.

3. Unlike many parts of ecosystems, insect-plant relationships have at least one sessile component (often two), making the study of their relationships very con\'enient.

4. Insects and plants have evolved in response to each other, and scientists are curious about how these relationships developed.

5. Insects and plants together comprise perhaps three- quarters of the known species on Earth, and the specific insects that feed on plznts may represent one

Insects and Their Environment: Plants

During the last 100 years of research on insects and plants. science has advanced enormously. During the mid- eighteenth century, Dar\vin estiinated that approsimatelg SOO.OOO species of organisms existed on Eanh. This esti- mate ivas raised to 1 million in the mid-1900s by >letcalf (1940). Even as late as the early 19POs, it ulas still esti- mated that only 1 million species of insects inhabited the earth (Strong et al. 1984), although other studies raised this estiinate to as high as 10 million (Gaston 1991). Recent re- search in tropical rain forest canopies has led to a drastic revision of these estimates. Only 10 years later, in 1992, we now think that there may be as many as 30 million in- sects, a 30-fold increase (Erwin 1982: Wilson 1988, but see Gaston 1991). Mort of that increase appears to be a consequence of insect communities in forests (both in trees and in soils), which may be much more complex and abun- dant than previously thought. Wilson has even conjectured . that as many as 100 million organisms could exist, based on the fact that s o much taxonomic work remains to be completed in many ecosystems, especially tropical rain forest canopies (\Vilson 1992).

In this section, we look at the ecological relationships of insects and plants, comparing the abundance of insects on structurally simple plants and \vorkin,o up to the more .:

cornplex distributions of insects in forests. Insects are associated with plants for a number of rea- :j

sons. The most imponant relationship involves food SUP- '<

ply: many insects on plants are herbivorous or phy- :.$ 2

tophagous (pliyto = plant, p l~agol is = feeding). Others, :: o Ira\'- .: however, use plants for territory, protection, matin,. .i

eling, communal roosting, and other activities (see also 4 :2

chapter 8, "Insect Behavior"). The colonization of habitats by arthropods and their .:

active (and sometimes passive) abilities to colonize a site have been the subject of extensive research in the last : decade. It is assumed that most arthropods colonize a hab!-

- Fip re 10.1 T h e energy pyramid of a land-based comnuniry. Ncltc that a large proportion of the energy is lost benveen t r o ~ h i c levels. Trophic le\~els are pictured and labeled o n the left.

Calories oer vear

Secbndary carnivores (tertiary consumers)

(secondiry consumers)

(primary consumers)

Figure 10.2 The number and proportions o i species in major tasa. escluding fungi, algae, and microbes. T h e proportion of phytophages is bared o n estimates for British insects, assuming thzt p r o ~ r t i o n s in the \vorld as a whole are hroadly rimilar. From D. R. Srror.2, J. J. Loudoun, a d R. Southu,ool, Insects on Plants. CczT;: Q !9S4 Blorl;wz!! Science Ld.. O~ford, England. Rem'nrtd

pcmission.

and that species will coexist if other populations do not in- hibit them or outcompete them. This perhaps originates from the original views of a plant ecologist named Gleason (1926), who advocated that communities are comprised of individual populations, with a coniinuum of species along an environmental or habitat gradient. Gleason's definition of a community is in contrast to that of another biologist, Clements (1916), who defined communities as "organisms" comprised of highly interdependent populations. The indi- vidualistic view of Gleason and the organismic view of Clenlents both have come in and out of fashion over time. but they draw attention to the importance of populations and their interactions in any community.

Populations of insects in communities fluctuate more extensively than do populations of longer-lived or~anisms. These oscillations have both positive and negative effects on field studies of insects: insects are easier to study be- cause several generations and population cycles can occur during one season, but population dynamics may be more difficult to interpret because numbers may vary widely over a relatively short period of time. Functional groups of insects in a community exist at two levels: populations of a species, and guilds of similar species that perform the same role (e.g., sap-sucking insects).

For purposes of study, one plant can coinprire a com- munity with its resident populations of insects, whereas one ecosystem can also comprise a conlmunity, ivith many plants housing many populations of insects. For our purposes, an illsect comn~unity is defined as populations of insects that interact on or in the vicinity of a common host plant.

The reasons for differences in numbers of both p o p - lations and guilds of insects on plants are many. \Ve dis- cuss three in this chapter: structural complexity and age of host plants, nutrition and toxicity of plant tissues, and evo- lution of mutualistic relationships. The topic of insects and plants is sunlmarized by a discussion of forests as the most complex communities where insects and plants interzct. Recent studies in forest canopies suggest that many as- pects of insect-plant interactions remain to be studied in this relatively unstudied "frontier region" of science.

Structural Complexity and Age of Host Plants

Insects on Structurally Simple Plants It is \\fell established that plant-dominated ecosystems have more insects than landscapes that do not. For example? volcanic peaks and alpine aeolian systems (i.e., regions that depend upon wind to import both energy and nutri- ents) have very sparse arthropod f2unas (Edwards 1987). As areas develop soils and then vegetation (termed succes- sioii, a classic sequence whereby a region underzoes 6.- irelopment of soils and vegetation, often from bare soil to field to shrubs to scrub to forest), a larger variety of plants is established, followed by a greater fauna of insects. As succession of the landscape leads to greater complexity, i t

is hypothesized that increased complexity of plants results in greater insect fauna. The numbers of insects also in- crease as plant size increases: from monocors to weeds to perennial herbs to shrubs to trees (Lawton 1983).

Mor~oclrltlrres (single-dominant groups of a species, as opposed to polyculture) of cereal grains o r other crop plants are both structurally and taxonomically simple, but their resident populations of insects have extremely'com- plex behaviors (see Liss et al. 1986). On the one hand, populations of insects in monocultures tend to be taxo- nomically less diverse than insect communities in more di- verse habitats; on the other hand, their populations may fluctuate enormously with subtle changes in the quality and availability of food. Price (1976) examined seasonal changes in arthropods on soybean plants, with predictable increases in numbers as the season progressed, followed by decreases in numbers as the plants underwent senes- cence. The number of species fluctuated from 0 to 30 dur- ing a three-month period. Bach (1980) examined the popu- lation dynamics of insects on cucunlbers in monocultures and in polycultures. She found that the populations of striped cucumber beetles (iicalji71i,~a ~ i t t a t u n ~ ) were higher in monocultures than in polycultures (figure 10.3). This example supports the resource concentr~atioi~ h~pothesis (Root 1973), which predicts that insect herbivores u.ill reach larger per-plant populations in areas with high densi- ties of host plants than in regions where host plants are patchy or rare. The enenlies I~ypothesis is another relevant idea, which predicts that predators and parasitoids should control numbers of insects more effectively in mixed stands than in pure stands.

Both of these hypotheses are pertinent in recent experi- mental ecology. The relationships between insects and crop plants are studied extensively in a discipline called inte- grated pest management (IPM; see chapter 17). This topic is of obvious importance to the sunii'al of human beings because we depend upon crop plants for our nutrition and rry to minimize the losses of these crops (cereals and oth- ers) to insect epidemics. Because crop plants are usually herbaceous, grown in monocultures, and structurally sim- ple, it is a challenge to agriculturalists who need to protect the crops from excessive increases in insect populations.

Insects on Structurally Complex Plants Lan.ton (1983) further hypothesized that plants offering a ,oreater variety of resources (e.g., more leaves or branches) u80uld support more individuals and more species of in- sects. Very little experimental evidence exists to support this hypothesis, however, because i t is difficult to test. In one experiment, Bach (1981) found higher numbers of spe- cialist chrysonlelid beetles on vertically grown (versus hor- izontal) cucumbers. This suggests that insects prefer the structurally vertical plant with its more complex arrange- ment of stems. With larger populations, species are more likely to persist on a plant despite predation, inclement

Figure 10.3 Mean number of sniped cucumber beetles (Acalymrrz r~cura) on cucumber plants (Cucumis sadws) in garden plou in monoculrures, or in polycultures planted with corn (Zen mays) and broccoli (Bracsia oleroceo). Density of cucumber plants (269 or 144/1@ ml), by itself, has little influence on beetle density. From 'Eflccu of Plant Dmtity and Diuersir). on the P o G d o n &M& ofa Si~abLt Herbime, the Smpd CYN& Bee&, A+ynmcl M'rw (Fob.)" by C. E. B d , Ecology, 1980,61,1515-1530. Copyight 0 1980 by r k Ecobgkd Society ofAmnica. Reprind bypmm'ssion. '

- High-density monoculture + - - -a Low-density polyculture

.90 - c- - -+ High-density polyculture

54 70 90 110 130 150 Days ~i.5: planting

weather, and so forth. Following Lawton's (1983) original progression, shrubs, saplings, and ground vegetation are slightly more structurally complex than low-grouting, herbaceous monocultures.

Some studies indicate that numbers of individuals (not just species) also vary with incresing structural complexity of the host plant. Saplings of eucalypts and closely related angophoras in Australia were examined for abundance of arthropods over time, and it was found that numbers fluctu- ated seasonally from 0 to 20,000 (figure 10.4). Some of these insects (e.g., Diptera and other flying insects) ap- peared to be situated on a plant as a result of chance, as travelers; others, such as the ants and aphids, were residents and also very prolific. This combination of transient and permanent residents on a plant accounted for the high vari- ability in numbers of insects on the eucalypt saplings. Other factors such as foliage quality and pheilology (seasonal pzt- terns of growth-related activities) of the leaves affected the numbers on a seasonal basis. These numbers appear to be extremely high, but eucalypts seem adapted to relatively large insect populations as compared to their northern tem- perate forest counterparts (see the section below on forests). Mature eucalypts have been observed t o have many tens of thousands of Christmas beetles (Anoplog- narltes, Scarabaeidae) feeding in their canopies (Heatwole and Lowman 1986, Lowman and Heatulole 1992).

Studies comparing the abundance of insects on seedlings, saplings, and adult trees are rare because of the

herbivory, specifically on feeding damage, which is an in- direct measure of the numbers of insects on plants. Coley (1983) examined the herbivory of rain forest saplings in Panama. Persistent, late successional species had greater defenses of leaf material and subsequently lower numbers of herbivores per tree than pioneer or early successional species of saplings. Similar results were obtained for the herbivory of mature rain forest canopy trees in Australia: the gap-colonizing giant stinging tree (Dendrocrtide ex- c e l s ~ ) had higher herbivory than the late successional red cedar (Too~la a~csrralis) and sassafras (Doryphora sas- zafras) (329,5%, and 13% annual leaf area losses, respec- tively) (Lowman 1985).

According to the structural complexity hypothesis, trees &id forest communities house the most abundant and diverse insect populations. In his classic work, Southwood (1961; see also Southwood et al. 1982a, b) made extensive measure- ments of the arthropod fauna of various deciduous trees. He used an insecticide spray which harvested all insects within the canopy and essentially provided a "snapshot" of each tree at one point in time. This technique, called fogging, is useful for obtaining a comprehensive count of all insects within a complex structure such as a tree canopy. Comparisons zmong trees and continents showed a fairly consistent pro- prtion of different trophic levels of insects: onequarter phy- tophagous by species but over one-half by weight, one- quvter parasitoids by species but less than one-sixteenth by ueight. one-quarter predaton by species and the same pro-

time and energy they require. It is possible, however, to p i o n by weight, on '

Figure 10.4 Seasonal changes in n u m k n of insecrs on 10 individual eucalypti saplings growing in close proximiry in an old field ne2 Amidale, NSW, Australia. Arthropods responsible for peaks include anu (A); psyllids (P); Christmas beetles (C); leafhoppen (L). Sapling species are Ango?howbuda (0, b); E d y p t ~ cBgimso (c, d); E. blakelyi ( t , f); E. meUiodoro (g, b); and E. uiminalis ( i , j). Darn born Hcorwo!~. Lcum, Donovan, and hfcCqv.

10.5 Composition of rhe arrhropd fauna of seer in terms of major guilds, expressed as mean number of sFeciu, number of and biomass, based on comparable samples from six specier of nee in Britain and six in South Africa. C - chewers,

s . sap-sucken, E = epiphyte fauna, Sc = sc2\.enging (dead wood, and so on) fauna, PR = predators, P -. parasiroids, A - mu, T - touris=. from D. R. Snofig, J. I. bwdoun, end R. Smhuw!, Insects on Plants. Cop)n'8fu Q 1984 Bkukwcll Scimce Ld., Odd, En&nd. Rcpimd pnhsion.

Species

Individuals Biomass

:e), epiphyte fauna, and ants (figure 10.5). In total, they und bet\veen 180 and 425 species of insects for Sa lk olba ~d Querclcs canopies, respectively, using a composite of )th fogging and observation methods.

More recent studies of insect abundance and diversity tropical forests may provide some of the most important formation about insect populations on structurally com- :X plant communities (reviewed in Lowmin and Moffett '93, Lowman and Nadkarni 1995). Tropical forest ~ o p i e s contain the highest structural complexity of any

trees, and emergents (in approximate order of increasing size) senle as variable hosts for insects. The functions of these host plants vzry but include shelter, food, or tempo- rary roosting for insects (see also chapter 8 on behavior).

The technique of fogging has been the catalyst for studies of insect biodiversity in tropical rain forest canopies. This sampling technique harvests the diverse in- sects from tall tree crowns, which may otherwise go unde- tected by scientists. The diversity and abundance of insects in tropical tree crowns appears enormous. Data from dif-

and unpublished); Brazil (Erwin 1995); Panama (Erwin 1995); Peru (Erwin 1982; David Xickle, personal commu- nication); Australia (Basset 1988, Kitching et al. 1993 and unpublished); South Africa (Southwood and Kennedy 1983); and Costa Rica (Longino 1991). Because the fog- ging technique produces such a large sample of insects, the analyses of the catch can take months and even years to as- sess, resulting in a potentially enormous taxonomic back- log for the entomologists in\lolved.

than younger ones (South\vood et al. 1982a, b). South- \\*ood (1961) compared'the numbers of insects between dominant native trees and introduced trees in Britain, Swe- den, Russia, and Cyprus. They found that the older tree comnlunity averaged higher numbers of residsrt insects, as compared to introduced, less \+'ell-established tree species (presumably where the trees and insects have had less time to reach an equilibrium).

Plant Nutrition and Herbivory

Insects on Plants of Different Geological Ages Plants are the largest and most readily a\lailable source of food in terrestrial communities (see Bernays and Chapman

It is assumed that insects have coevolved with plants for a 1994). Phytophagous insects have the capacity under idea] long time, so the older, more complex and more highly conditions to increase by a geometric progression (expo- e\lolved ecosystenis may have richer insect faunas and nential]~), but this capacity is rarely realized in nature. In more complex relationships between plants and insects. an influential paper, Hairston et al. (1960) obsened that Once again, these hypotheses are difficult 10 test because plants seem to be damaged extensively only \\hen nanrral they involve e\rOlutiOnarj' time, which cannot be manipu- enenlies (parasitoids and predators) of herbivores are ab- lated experimentally. Biologists can look at different mem- sent, due, for instance. to the greater sensitivity of enemies bers of one group, houVe\'er, and compare the differences to pesticides or the introduction of herbivores without their of the plant-insect relstionships they obser1.e. enemies. F ro~n this the! deduced that herbivore populations

One well-studied group of plznts are the thistles. Eu- are conuol]ed by their natural enemies rather than by avail- rope, especially the Mediterranean region, has a large and ability of food, in contrast to plants and carnivores, \\.hich ancient thistle flora used by more than 50 genera of insects. they =gued conuolled by conlpttition for resources. hliost of these species are host-specific (feed just on this- In some cases predation, parasitism, and disease con- tles) and an especially large guild feeds inside [he flower trol the abundance of herbivores, but experiments to test heads. One tribe of thistles migrated to h'0d-i America and these regulatory effects are difficult to conduct (Price speciated extensively. But the diversity of insects on these 1987, price et 1990). w h y [hEn is the world green? thistles is low, apparently because few thistle-feeding t aw Hairston et a]. (1960) assumed that anything green is food accompanied their ancestors as they migrated from Europe for herbivores, just as different antelope species are over the Bering Strait land connection in the latg hlIiocene equally good for large predators. we shall see, [his as- (Zwolfer 1988). sumption has proved false.

A perhaps older and more complex \veb of relation- Herbivorous insect species are very abundant, com- ships exists between Heliconius butterflies and plants in prising perhaps half of all individuals found in nature. neotropical rain forzsts. These butterflies have a variety of Holr.ever, out of 32 orders of insects, only 9 feed on living .. interactions with their sessile hosts; adults are brightly col- plants (figure 10.7). nis suggests [hat plants are a foh- .j ored, a warning to \\'auld-be predators that they retain the dable e\rolutionary barrier that most groups have not been toxins their lXr~ae obtained from their passion vine hosts. able to o\,ercome (Bernays and Chapman 1991). But the 'T

:. . Warning is enhanced by greg2riouS feeding of larvae? high diversity of insect species \,,:ithin these orders sug- -.; ." communal roosting of adults and participation of adults in zests that, once these barriers are surmounted, insect ,? mimetic complexes \t1ith other distasteful butterfly species ;,ups radiate extensively (h{itter et al. 1988). (there are two kinds of mimics: Batesian mimics are palat- able species that look like distasteful or dangerous model

visit the same plants repeatedly) by following older adults chapter 4), so they must mechanically disrupt or othewise

1981 on tea trees) and that older forests have a richer fauna, plants for food.

Figure 10.7 The sire of the insect orders containing significant numbers of herbivores. The s t e ofhe order tokmfrm Bmor et d . . 1976 a$ Richad and Dab%, 1917. Proponion ofherbivores in each order taken from Price, 1977, txcept for Thysanopura, which was a d d from Ltub, 1973; from Weis and B t r d u - m , in Plant-Animal Interactions, W Abrabnson, td. , 1989. From W . G. A b r h m o n , d., Plant-.b.ir.r! Interactions. Co~,n'ghc 0 1 9 s >f&raw-Hill, lru., New Ymk, A?'. Repimd by pzrission ofhfcGraw-Hill, lru.

Number of species (x 1 05) 0 1 2 3

Coleoptera Beetles

Hyrnencpiera Sawflies, bees, wasps

Dipteri True fiiss

Hemipiera True bugs and Hornoptera aphids end relatives

Orthoptera Grasshoppers

Thysanoptera Thrips

Phasrnida Stick insects I

Herbivores y&+ Others

Chewers cause most of the material eaten by chewing insects

Chewing is the most common way in which insects cludes tough cell walls, the mandibular surfaces are

process plant material, including leaves, stems, flowers, posed to considerable wear and are replaced at each molt

pollen, seeds, and roots. In the orders Orthoptera, Coleoprera. Hymenoptera, and Phasmida both juvenile and Miners and adult stages have chewing mouthparts. In the Diptera and Larvae of many insects and some adult beetles feed wi

more of the tissue layers beween the intact upper and lower of leaves. As the insect tunnels its way through a

leaf, it creates a mine whose pattern is often characteristic of the species. Except in the Diptera, miners tend to be flattened to the dimensions of the leaf. Like other in-

that feed inside growing tissues, miners often feed on one or a small number of related plant species. Leaf-mining insects are found in the Lepidoptera, Diptera, Hymenoptera, and Coleoptera. The greatest diversity of forms and number of leaf-mining species are in the Lepidoptera, followed by b e fly family Agromyzidae (Hespenheide 1991).

Boritig insects live in the woody tissues of plants. The abundance and long-term availability of food provided by tree trunks permits growth to a large size. The larvae of Syleutes boisduvale (order Lepidoptera, family Cossidae), for instance, attain a length of 18 cm (Waterhouse 1970). In addition, the low nutritive value of wood tends to result in long life cycles. As an aid in using \vood for food, some borers inoculate the \vood with fungi that attack the plant and upon which the larvae feed. Dutch elm disease is caused by a xylem-blocking fungus inrroduced by the elm bark beetle, Scolytw scoljtus. Other borers attack the liv- ing cambium and vasculv tissues, which can devastate the host. Bark beetles in the family Scolytidae are examples (Mitton and Sturgeon 1982).

Gall-Forming Insects Many phytophagous insects induce the production of abnor- mal growths, galls, in the tissues of their host plants, hiside which they live and feed. Galls are found in buds, leaves, stems, flowers, or roots, and their s h a ~ s and locations are often characteristic of the plant and insect species concerned (figure 10.8 a+. In galls initiated by species with chewing larvae, the insect resides in an inner chamber lined by nutri- tive tissue upon which the insect feeds. The gall is entirely a product of the plant, de\doping in response to a chemical stimulus from the secretions of the insect. The diversity of gall types and the biology of host and gdl can be explored in books by Meyer (1987) and Ananthakrishnan (1984).

Galltnakers (also called gallers) reach their peak of di- \.ersiry on dicots, the broad-leaved flowering plants. In North America and Europe, 85% of gall-forming species are found on the oak (especially Quercus), rose, and sunflower families. In South America, galls are es~c ia l ly common on legumes, and in ~usrralia half of all known gallmakers are found on plants in the Myrtaceae, especially on Eucalyptus trees (Meyer 1987). Gallers are found primarily among the Diptera (Cecidomyiidae) and Hymenoptera (Cynipidae), but there are some representatives among the Hymenoptera

the plant epidermis and suck plant fluids. In this way, true bugs (Hemiptera), aphids and their relatives (Homoptera), and Thysanoptera avoid consuming indigestible structural components of the plant as well as toxic materials pro- duced by the plant.

Sucking insects sunrive on several types of plant flu- ids. Thrips pierce and macerate the contents of individual cells with their stylets and then suck the liquified material through their proboscis. Other species (e.g., the meadow spittlebug; 13- and 17-year periodical cicadas) tap xylem \vessels, from which they get a very dilute nutrient solu- tion. But most hemipteran species insert their beaks di- rectly into phloem cells to tap the more nutritious fluid there. Phloem fluid is a sugar solution with low concentra- tions of other necessary nutrients. The specialized filter chamber digestive system that allows phloem feeders to make use of this fluid is described in chapter 4.

The feeding of Hemiptera and Homoptera invol\-es neg- ligble mechanical injury to plant tissues. However, saliva in- jected by some species may be toxic, and sucking mouth- parts are often invol\?ed in the transmission of v i n l diseases. So some insects are of more importance as plant-disease vec- ton than for the direct damage they inflict by f d i n g .

Seed-Eaters Seed-eaters (and seedling-eaters) are the only true plant predators among insects because they kill plants by consum- ing them. Herbivores that consume stems, leaves, buds, or fluids are likened to parasites because the host is not killed (Pnce 1980); a growth module of the plant (either branch or leaf) may be killed, but the plant can, grow a replacement pm. Seeds are a rich source of protein and are used by both sucking and chewing insects. Seed beetles (Bruchidae) are one of the largest groups infesting seeds. Janzen's (1971a) studies of this group in the neotropics led to the development of some important ideas in ecolog (e.g., predator satiation).

Demnvores Although greatly overlooked until recently (see the review by Hoover and Crossley 1995), insecis that fezd on senes- cent or dead plant parts comprise a very important part of most ecosystems. These insects usually live in the soil, and because of the difficulties of counting or e\*en locating in- sects underground, they have been o\.erlooked in most ecological studies of insects. Detritivores include some Grylloblattodea and Blattaria (cockroaches), some Coleoptera (beetles), Isoptera (termites), Psocoptera (bark and book lice), and Collembola (springtails).

(Chalcidoidea, Agaonidae), ~ o l e o ~ t e r i Thysa;optera; and Lepidoptera, and in the fluid-feeding Hornoptera (Aphidi- Plants As Food dae, ~ ~ i o m o r ~ h i n a e ) (Waterhouse 1970).

Plants, permanently rooted to a spot, cannot flee from their consumers. Therefore, it is easy to assume that plants are

Sucking Insects

. - -- Firmre 10.8 Insect galls. (a) Andriw partmow (Hymenoptera; Cpipidae) on leaf of blue oak. (b) Pine-cone willow gall caused by R h h p b g a snobiloidk (Diptera; Cecidomyiidae). ( c ) P m p h i p beta (Homoptera; Aphididae) on Populur anpdfolia. (dl Longitudinal section through the ball gall caused by Eurosta solidaginis (Diptera; Tephritidae). .. . - (a) INSECTS AND MITES OF WESTERN NORTH AMERICA by Essig, E. O., Q 1958. R ~ p ~ n r e d bj jxtmksion of Prmdce-Hd, lnc., U p p S& R i m , N). (b) Repinred by pemissionfrom G m d Biologicd, lnc., Chicago, IL. (c) From 'Habirat Sekcdon by Pemphigur Aphids in Response u, Resource Limitadon and Compeddon" b-j T. G. W h m , Ecology, 1978,59, 1 164-1 176. Cohn'ghr Q 1978 by h e Ecologicd Society of A m ' c a . Reprind by permission. (d) From W . G. Abrahamon, ed., Plant-Animal Interactions. Coa+ght 0 1989 McGrow-Hill, Inc., New York, AY. Reprindby pomission of McGraw-HiU, lnc.

Larval chamber

Nutritive tissue

Parenchyma tissue Stem vascular tissue Stem epidermis

different ways of feeding. Plants are only marginally nutri- and in crop plants. Some of these characteristics, including tious, and most plants have a variety of both toxic and structural materials, nitrogen, water, chemical, and physi- physical defenses aeainst consumers. Further, the effects cal defenses, are described below. of air pollution on plant quality may exert indirect effects on phytophagous insects (see boxed reading 10.1). The Snuctura/ Materials wide variety of characteristics plants use to defend again:[ The rigidity of plants is imparted by the structural carbohy- herbivory have been the subject of extensive and exciting drates cellulose (see fig. 2.2) and hernicellulose. As plant tis-

- B O X E D R E A D I N G 1 0 . 1

Pollution-Induced Changer in Plant Quality

. n e r e is growing evidence that air pollution changes insect peifonnance and levels of attack on plants (Riemer and Whittaker 1989). At the broadest level, we might suspect that the nitrogen compounds in air pollution and increased C02 would affect ]el.els of structural carbohydrates and the concentration and composition of leaf proteins. Air pollution has resulted in dramatic inputs of nitrogen to the environment over the last 30 years. Emissions of NO2 come largely from fossil fuel combustion and NH, from fertilirer applicario~~s (xhich have quadrupled since 1945) and intensive li\~estock operations (Wedin and Tilman 1992). Insect responses to fertilizers and to pollution.stressed plants are variable, and we are not now able to predict how these factors alone, much less together, \\.ill influence insect populations.

The consequences of increasing atmospheric C02 levels on interactions between plants and their herbivores have been in a small number of studies. Some plants grown at high C 0 2 levels had lower nitrogen levels, presumably because of

dilution by cellulose and other carbon-based compounds, and insects increased their consumption of high C02 plants by 20% to 80%. Despite this, they grew more slo\\~ly than larvae on plznrs gro\\.n at ambient COl levels (Bazzaz 1990). In contrast, sagebrush grown a t high COz \vas nore readily digested by grrsihoppers (Johnson and Lincoln 1991). Clezrly, the direct and indirect effects of air pllurion are nor going to be easy to predict.

toughtr and morz difticult for insects to che\il or pierce. In studies of Australian rain forest leaves. herbivory of young leaves \vas 15-40 times s t a t e r than that of old leaves, znd leaf toughness increased similarly (Lo\vman and Fox 1983). There is a large amount of energy bound up in cellulose and hernicellulose, but insects cannot digest them since they do not produce cellulase, which is necessary to brcak down cel- lulose. The bulk of what h o s t all cheaing insects eat is con- sequently indigestible. Important exceptions to this are insects such as termites, leaf-cutting ants, bark beetles, and cryptocer- cid cockroaches, jilhich have internal or external mutualistic associations with microbes or fungi that can produce cellulase (see chapter 4). Plants also produce \\.axes, cuticle, and other indigestible structural materials and concentrations of these, in addition to cellulose and lignin, increase as tissues age. They are especially high in tissues of plants go\ving in high light, dry, or nutrient-poor habitats where the vegetation is re- ferred to as scleroyhjllous (hard leaf) (Morrow 1983).

Insects are more than 50% protein and 7% to 14% nitrogen by dry mass because protein is a major component of their structural compounds (e.g., arthropodin, sclerotin; see chapter 2). In contrast, plants have nitrogen levels that are much lo\i~er, from 0.5% to 8% nitrogen (figure 10.9), so in- Sects must concentrate large amounts of protein from a very dilute source. As the nitrogen content of their food increases, insects become more efficient at extracting it (figure 10.10~). Many studies demonstrate increased growth (figure 10. lob), survival, fecundiry, and longevity with increases in nitrogen. It would appear to be adilantqeous for insects to maximize their consumption of this limiting resource.

One way for insects to increase nitrogen intake is to

leaves must have at least 1% nitrogen in order for lanfae of the chrysomelid beetle Paropsis aroi~~ar ia to gain more ni- trogen than they excrete and egest in their frass (feces; Fox m d Macauley 1977). For grasshoppers and larvae of Pro- delztia eridullia, the break-even level of plant nitrosen is 370 (Mattson 1980). But nitrogen content is more complex than nitrogen concentration alone implies. For exzmple, when nitrogen levels are elevated by the application of fer- tilizer, insects do not benefit as predicted from studies of plants that have naturally high nitrogen levels or levels ele- vated as a result of stress (Morrow 1983).

The relative proportions of amino acids that make up plant and insect proteins are different. In artificial diets, Karowe and Martin (1989) found that the respiration rate of Spodoptera eridania larvae depended upon the amino acid composition, but not upon the quantity of protein. Southwood (! 973) concluded that predatory insects con- vert animal proteins to biomass more efficiently than her- bivores convert plant protein. One major difference is in the ratio of phenylalanine to other amino acids. Plants pro- duce relatively small amounts, whereas insects require large amounts to make chitin. Grasshoppers, for example, have significantly increased growth rates when phenylala- nine is added to their diet (Bemays 1982).

T h e nitrogen content of plants varies considerably m o n g plant species and tissues (figure 10.9), and with the season. Actively growing tissues require high levels of ni- trogen to support protein synthesis, but once the tissue is fully expanded, synthesis shifts to suuctural carbohydrates, 2nd nitrogen levels become diluted. The deciduous oak, Quercus robur, has leaf-nitrogen levels of about 5% at bud burst in late April, 2.5% nitrogen when fully expanded, and 1.5% just before leaves are shed in October (Scriber and Slansky 1981). The equivalent values for Eucalyptus pauci-

Figure 10.9 Variation in nitrogen content of animals, fungi, and difierent plant tissues. Bars span the range of typical values. ~ f v t Strongcrd., 1984.

Nkrogen content (% dry mass)

Figure 10.10 (a) Efficiency of con\-enion (biomass gainedJbiornss consumed) for various invertebrate phytophages as a function of the nitrogen content of their food. Lou.-nitrogen diets are used inefficie~tly because the rate of feeding and of throughput in the gut are increased to compensate for low food quality; therefore, digestion is less efficient. (b) Effect of leaf nitrogen content on development rate of Paropsis atomnria lanae (Coleoptera; Chrysorneli?ze). The number of days from hatching until most larvae in a cohort become prepupae decreases as nitrogen content increzses. Percentage nitrogen x 6.25 gives an estimate of protein content. (a) Source: W. hlartcon, 'Herbiwry in Rcladon ro Planc Nitrogen Conur.~' in Ar-~ual Review of Ecology and Systcrnatiu, I 1 :I 19-162, i980, Ann& Revieuss Inc., Pdo A h , CA. (b) From P. A. Slmw and L. A. Fox, 'Ej'tctr cj\';?adon in Eucalyprus Essendal Oil Yield on Imect Grouch a d Herbivore bad" in Oe-ologia, 45:2CY-219. 1980. Coz+ght O 1980 Springer-Vcrk:Stu. \'A, Inc., h'ew York, KY. Repined by p-mnission.

" 1 2 3 4 5 6 7 8

I- \ Nitrogen content (%) of plant tissues .. . 1 2 Nitroaen content (%I of leaves

Nitrogen levels in phloem fluids also vary. When the ,]ant is actively growing and later when nitrogen is being re- ,-je\led from leaves before they are dropped, the concentra- ion of nitrogen msported in the phloem may reach 0.6%.

between these periods, le\'els may drop to .03%, and the ,roprtions of amino acids change mattson 1980). Phloem- apping insects such as the green spruce aphid, Elatobiurn ;bierir~~irfr, respond to this nitrogen change by producing ,.inged individuals that disperse from the plmt in late spring .O find alternatiI1e hosts OlcNeill and Southeood 1978).

n e water content of foliage varies from 45% to 95% of fresh weight. All else being equal, leaves \vith high water content (e.g., expanding leaves or leaves of annual plants) are digested most efficiently and support h e highest rates of growth (Scriber and Slansky 1981). Lezf water content is correlated with nitrogen content, and both decrease with increasing levels of structural materials.

Adaptations have e\.olved in higher plants that enable them to deal \vith the harsh realities of extracting water and nutrients from dry land. The specialized \rfays of plant feeding that have evolved in phptophagous insects provide evidence that plants have posed difficult nutritional prob- lems for insects to'overcome.

Insects that have evolved the ability to use particular host plants exert selection pressures that favor the evolu- tion of defenses in those hosts. We will consider these co- e\rolutionary interactions later. Here we simply outline the main characteristics of plant defenses.

Plant Defenses Plant defenses against phytophagous orgznisms of all sorts fall into two basic categories: physical defenses and chem- ical defenses that are toxic or deterrent, or that actively in- terfere with digestion.

Phgsical Defenses Plants commonly have tough surfaces thzt make them hard to pierce, bite, and chew. Toughness is endowed by thick cell walls with high levels of lignin and by deposits of sur- face waxes. Silica crystals are common in epidermal cells of grasses and horsetails, Equiserum, and abrade the mandibles of herbivores.

Many plants are covered by small hairs called fri- clto~nes, or pubesce~lce. These hairs come in many different forms: long, short, thin, sharp, flattened, slippery, and with or rrfithout glands (Levin 1973). Leaves md stems of Passi- flora adenopoda are covered with small hooked hairs that protect it from larvae of Heliconius, u-hich are specialized feeders on passion vines. Where hooked hairs pierce the larva's skin, hernolymph drains out, and the larva dries up

In contrast, the leaves of the giant stinging tree (Ur- ticaceae, Dendrocl~ide excelsa) in tropical Australian rain forests are densely covered with stinging hairs, as the name implies. Observations by Lowman (1985) have shown that a host-specific beetle, Hoplosrinis ~jiridipenis, is undeterred by these defenses. Tine leaves lost an average of 32% leaf surface area during a lifetime. Stinging hairs may have evolved as a defense against mammalian herbivores in Asia, but are less effective against insects (figure 10.1 lc). Defenses may be ineffective against some herbivores, but if they deter other potential herbivores, then the plant has won at least those battles.

Chemical Defenses Plants are the original pharmacies. Among the generally \\ell-kno\vn compounds they produce are vanilla, salicylic acid (aspirin), caff<ine, nicotine, morphine, pyrethrum, and tannins. Fraenkel (1959) argued convincingly that the di- verse array of compounds produced by plants u e defenses against herbivores. In other \\fords, plants have evolved to produce chemicals that deter consumption of their foliage and orher tissues. An enormous literature on relationships bcnr.scn plant toxins and herbivores now exists that sup- ports Fraenkel's view. This literature is rcvie\ved in the works of Rosenthal and Janzen (1979), Rosenthal and Bertnbaum (1991). and Harborne (1988), and is published in journals such as Cliemical Ecology.

Defensive chemicals, also called allelochemicals, or secondary co~npounds, are produced from biosynthetic pathways related to the primary metabolic pathways (figure 10.12) common to all or large numbers of orzan- isms. Plants produce an incredibly diverse array of these compounds, and all of them can be traced back to a very limited number of precursors, acetyl coenzyme A, meval- onic acid, and shikimic acid (Waterman and Mole 1989). Allelochemicals can poison the physiological systems of herbivores, many of which are the same as those of plants. Plants avoid self-poisoning by isolating allelochenlicals from physiologically active sites, for instance in glands. resin ducts, or vacuoles. Such chemicals may also be stored as inactive precursors that become damaging only \\hen metabolized by the herbivore or when the plant is dzmaged. For instance, free hydrogen cyanide, which poi- sons mitochondria1 respiration, is released from damaged clover leaves when cells are ruptured and cyanogenic gly- cosides and glycosidases come together (Jones 1973). Free cyanide can deter insects and is even more toxic to mam- malian herbivores (Bernays 1982). Over 100 phytoecdys- teroids alone have been isolated from plants; like other chemical components, they probably defend the tissue from herbivory (LaFont and Horn 1989).

We can roughly categorize allelochemicals into two groups: those that exacerbate the already low nutritional

and dies (figure 10.lla,b). Trichomes u e an important di- valud of plants and those that are toxic to basic metabolic

Fipre 10.11 Lanpae of Heliconisls rnt:j;:xanc e.?: ~;:a::y 5;ei i t~ i.i r'sriibjtora but not 17. acitnopoda {or the reason illustrated in these scanning electron micrographs. ( a ) Thir? instar I x r a i:ugl:t I.:] !t;ipc:iole (85x1. Proleg marked s is enlarged in (b). (b) Proleg has hooked trichomes embedded at a, b, and c iausir;; the i > i ~ ~ l skin ;o split and hemolyniPh to drain from the wound (Sjx). (c) Host- specific chrysomelid (Hoplosrinis dndipe:::r) feeds O:I tilt ;innt rri:?siag tree in Austrzlian rain forests, obviously undeterred by h e hain. Hairs have a sharp point that on contacr ?etaih, i :~ j tc r i :~s j y ~ i s ~ : : i \\.hose effects may t e felt by humans for up to a month. (a, b): From Giben, L. F . , Scknce: 172,565-5 I I?;; ) . 2 A.i.2.S. t i ! . L ?!. Lu7.m.

( t a b l e 10.1). Ea t ing becomes a rr.clrc r ime-conzun~i r . ; process and increases the length of iiillc zil irliect is c\.- posed to predators, parzsites, and er,\.ironl:iental haz2iC~. These chemicals generally are pres:nt i n I~i_rh c011ce11i::- [ions (e.g., up to 23% of the dry mas: of Siika .zpiucr: bri:.: is lignin, and 21% of eucalyptus I s ~ \ c s nl?y t?c cssefiiiki oils), and their effectiveness is dose-itpcndciit ( \ i ' n i n l ~ c ~ ; t et al. 1990). Quantitative compounds x c u m ~ l l n ~ c as tissct i age, and the plant does not scavengt their larpely cclrScr. components before shedding the stmciurr (hei~ce. i~irrrrol;i.'t. compounds). Compared to the next sroup of allclochc.r.-r.i. cals, digestion reducers are not very c ; l\.erse.

The second group of allelochemicals are referred to as iosic, qiici1i1oiii.e. or rlrobile rr;lrpour~ds. Qualitative com- pounds are present and effective in small concentrations (generally less than 5% of leaf d ry weight). They tarset es- stntial functions of herbivorous animals such as respira- tion, DXA repair, transmission o f nerve impulses, and hor- n o n e production (table 10.1). Often they are synthesized 2nd deeraded o\.er fairly short periods, hours to weeks, and their component pairs end up in other compounds. In the plant they may have nondefensive as well a s defensive functions. For instance, nicotine. a potent insecticide, also is used by tobacco plants to t ranspon nitrogen to the shoot

f ip re 10.12 The ent-kaurene to gibberellin (a hormonc) primary $athu.ay probably present in all higherplanrc. ~ i te&noidr indicated to the right of rhe dotted line are secondary metabolites characteristic of goldenrods, Solidago, which inhibit feeding by many

herbivores. From J. E. Graebe, 'Gibberellin Bios)nrhrcu and Cor.tol" in Annual Re\.iew of Plant Physiology, p g e 425, 1987. CoP-jigfit Q 1987 Annuol Reviews, Inc.,

p& .4ico, C.4. Reprrnted by pmnirsion. I I

//I7 I 2011~~~.13-16 " I I I I

2,'4'o'9\aE,5 * I I l l I 3,4,5,6,7 ent-kaurene I

18 I

a I I Powerful insect antifeedants

ent-kaurenol

ent-kaurenal 7P, 12a-hydroxy kaurenolide

ent-kaurenoic acid (

stimulant) COOH n I CO - -' co --'

a ent - 70. - hydroxy kaurenoic acid (

I

I I I

n COOH GA,,-aldehyde 1

COOH~ COOH

COOH snt-5a7a-dihydroxy keurenoic acid

T a b l e 10.1 Selected Plant Products that Reduce HerbivorJa1b

Chemical group (number identified) Description Defensive role

Quantitative Digestibility Reducers Cellulose (I basic type) Sugar polymer Requires gut flora for digestion Hemicellulose (1 bs ic type) Sugar polymer Requires gut flora for digestion Limpins (indefinite) Phenolic polymers Bind with proteins and carbohydrates Tannins (indefinite). Phenolic polymers Bind with proteins Silica (1 basic type) Inorganic crystals Indigestible

Qualitative Toxins Alkaloids (20,000) Heterocyclic N-containing Many; some stop DNA and RNA production Toxic amino acids (260) Analogues to protein amino acids Compete with protein amino acids Cyanogens (23+) Glycosides that release HCN Stop mitochondria1 respiration Glucosinolates (SO) N-containing K salts hfany; endocrine disorders Proteinase inhibitors (indefinite) Proteins or polypeptides in subunits Bind \\pith active site of enzymes Terpcnoids (1 00,000+) Polymers of Cf, units Many; some stop respiration

.See Rcren~hal and Janren (19i9) for a divuuion ofother secondarj c o m , ~ z ? s . bC, cuton: H, hydrogen: K, ptzs ium: Y, niuqcn; HCN, cyanide. Some compound, have chzaitcrirricr of bd; pOuFS. Silica, for htance , czn !x vitve.3 z> b:h a and a quantitari~e defense. 2nd rhe defensive ~roprrier of cetlulorc may be coinci3e?.:-.l to its primary ruppn function (Howc and B'u:lc!. !G55! . Source: Eenq F. H c ~ e and L!nn C \Verlcy. Ecolo~ccl re him ti?^ ofPlsr.rr L-2 .Z.-:%h. 1FE.S. Oxford Uni\.e:si:y Frelr, lnc.. S c w Yo:k. NY.

from uptake sites in rhe roots, a function performed in most plants by other nitrogen-rich but nontoxic com- pounds such as glutanline, asparagine, and amino acids closely related to them (Pate 1973).

Plant defenses may be constzntly present in a plant or they may be produced in response to tissue damage. Coristi- tutive deferlses are a plant's permanent protection. They in- clude most of the physical defenses like silica crystals and tricliomes as well as many digestion-reducing and toxic coinpounds (table 10.1). They help plants sunive first en- counters \vith herbivores. I~ldrtcible deferises are produced after the plant begins to experience damage (Tallamy and Raupp 1991; Haukioja 1991; Weins et al. 1991). The best- documented inducible allelochemjcals are proteinase in- hibitors. These polypeptides and proteins block the catalytic activity of digestive enzymes in the gut by binding to the ac- tive site of the enzyme molecule (Ryan 1979, 1983). Pro- teinase inhibitors are extremely common in the plant kin:- dom, including the legume family. They are denatured by heat, which is one of the reasons \re cook beans and some other vegetables. See chapter 4 for a brief discussion of hon genetic engineers hale used the proteinase inhibitor genes to develop plants that are resistaiit to cenain pest insects.

tion, transport, and storage of nutrients, and fluctuations in the concentration of toxins and other compounds that inter- fcre with feeding. Plant variation imposes severe con- straints on many insects (Whitham 1933; Schultz 1983).

The boundaries of those constraints have been studied for the aphid, Per71phiglts berae, by Whitham and his col- leagues. In the spring, females emerge from eggs that overwintered in bark fissures of their cotton\vood host tree, Populus ar~gusrifolia (figure 10.8~) . Galls started within three days of bud break have the greatest chance of being successful. By day seven, leaves are fully expanded and lignified and cannot be galled. Bud break is nearly simulta- neous throughout a tree, so there is little room for error in fe- male emergence time. This is not an easy problem to deal with because the date of bud break varies by 3-7 days among trees within a population (3-7 days) ( \hitham 1978) and for individuals from year to year (e.g., figure 10.13a).,

The gall aphid's temporal problem is exacerbated by spatial problems. On leaves less than 5 cni? in area, most 1 galls fail. Above this size, the number of offspring that emerge from the gall increases with leaf size, lvhich is re- lated to the amount of nitrogen the tree sends to a leaf; :: larger leaves get more than smaller ones. Most leal'es are 2 too small and the large leaves are scattered throughout lhe ;; - . .

Catchin: Food: Temporal and Spatial Availability crown. Females crawl from branch to branch searching for ;?: -3 unoccupied large leaves. The females actually fight one :;i ...

Plants u e dynamic. All of the pmpcnies so far discussed anatheifor pos~essian of the best leaves (Whitham 1978). :$ change with time and in response to biotic and abiotic envi- S o t unexpectedly, the number of galls on a tree varies ;2 rollmental challenges. The parts of a single plant can \xry enormously f r o ~ n year to p a r (hloran and Whitham 198s). .::$ enormously in their potential to support the growth and re- Trees in the large genus Eucolypnrs have a \'eQi differ- production of an insect. These differences have causes ent leafing ~henology. They initiate new leaves whenever it

7-

Figure 10.13 (a) Twenty-six year record of the date of bud burst for tagged branch of an aspen (PopuIus) tree. (b) Relative abundance and duration of leaf production on five saplings of Eucd~pw blokelyi at a single location in one year. The asynchrony in leaf initiation, which increases as the (southern hemisphere) grol~ing season progresses, is caused by individual responses of trees to different ler9els and timing of insect defoliation.

(a) Data from Hcdson. (b) Data of 1: 2. Fox.

April May

Diie of bud break

Leaf production

I I I I t I I I I Oct Nov Dec Jin Feb March April May

IJonth (b)

attack. This behavior can result in continuous groufth or in bursts of growth (figure 10.13b). Unlike the narrow window of time available for P. berae to exploit cottonwood, multi- ple large windows are available to eucalypt insects. This is probably one of the factors responsible for the high levels of damage that.eucalypts sustain in most years (Fox and hlor- row 1983, 1986; Morrow and Fox 1989). Furthermore, in- sect attack opens these windoxs wider because trees re- spond to anack by producing new flushes of growth, which the insects then eat. This is a major contributing factor in the eucalypt dieback in rural regions of Australia (Heatwole and Lowman 1986), a widespread tree decline that has ranked among the most severe in the world. Outbreaks of psyllids, beetles, and other insects repeatedly defoliate the canopies of eucalypts until mortality occurs. Defoliation can reach 275% per year, because the insects defoliate the canopy each timz it refoliates. The stress of repeated defoliation is all the more significant because the eucalypts already face other stresses, for exmple, widespread clearing of adjacent forest, soil compaction from li\*ertock, recurrent droughts, m d depletion of many insecti~~orous bird populations.

Recently, W i t h a m (1989) h2s shown that hybrid cot- ton\vood trees suppon several orders of magnitude more gall aphids than the parent species. A contributing fzctor may be the increased time over ivhich bud burst occurs in the hybrid zone. 21 days versus 3-7 days in the parent zonz (T. G. Whitham, unpublished data).

hiutualism and Coevolution Every organism must obtain enough food to rnsintain itself long enough to reproduce if it is to pass its genes to the next generation. As an organism proceeds from conception to reproductive maturity, it must constantly obtain food and at the same time avoid becoming food for some other or- ganism. Organisms that are unable to do this successfully, for whatever reason, are maladapted and leave few descen- dants. Although a number of en\.ironmental factors are un- doubtedly involved in the selection process. examination of specific adaptations often leads to the identification of spe- cific selecting agents. Such is the case when we examine adaptations associated with many interspecific interactions. Organisms that are closely associated ecologically often zct reciprocally as agents of selection; that is, they act zs se- lecting agents on one another. There is good evidence thzt many organisms have done so for long perifids of geologic time. These organisms are said to have coevolved.

Examination of several examples of coevolution re- veals two types of interactions. TWO different species or guilds of species that have coevolved to the extent that they benefit one another are termed t?rut~calistic. In this case, the presence of both enhances the reproductive suc- cess of both. Alternatively, two species may engage in a kind of genetic warfare; one adzpts a mechanism of of- fense that provides increased access to the other as food, only to have the other "respond," in the evolutionzry

- - -- - Figure 10.14 Bull's horn acacia branches (Asia comigera)

are occupied by an aggressive stinging ant (Pseudom)rmex fmginea). Ants live in the thorns and feed on starchy Beltian bodies and sugar-rich secretions from estrafloral nectaries. Th'e ants clear vegetation around host trees and attack insect and mammalian herbivores that might eat the foliage. From Ecological Relationships of Plants and Animals by H r n q F. Howc

and L>nn C. Wescky. Cohn'ghr 0 1990 O~jo;d L'niile~si~ Press, 1nc.

Reprinted bj: pennisrion.

prey, parasites and hosts, and herbivores and plants. For a comprehensive introduction to coevolution, see Thompson (1982, 1989). W e give two examples of mutualisms here: ants and plants and plants and their pollinators.

Ants and Plants Some species of ants in the neotropics live in acacia trees. An aggressive stinging ant (Pseudotnynnexferrugi~tea) nests in the hollow thorns of the bull's-horn acacia (Acacia contigera) and feeds on nutrient-rich Beltian bodies and sugar secretions from extrafloral nectaries present on all thz leaves. In return the ants attack herbivores and clear vegetation surrounding their acacia tree (figure 10.13). Acacias without ants have lower growth rates and are often killed by herbivores, fire, or crowding from other plants (Janzen 1966). In this case, the mutualism is obligate: nei- ther the ant nor the acacia sun.ives without the other.

More common than obligate mutualism is facultatire mutualism (i.e., the interaction occurs under some condi- tions but not others). One of the most common kinds of facultative mutualism occurs amon2 the many species of plants with extrafloral nectaries and the many ant species that seek nectar. While searching for nectaries, ants may prey upon herbivores they encounter. Both muwalists bene- fit from the interaction, but neither is dependent upon it . Tilman's (1978) study of ants, cherries, and tent caterpillus provides an excellent example of facultative mutualism.

Plants have a mutualistic relationship with many ants

4

Fipre 10.15 The phlox family (Pokmonixe~ceae) illustrates divergence of flower form due to selection from different pollinators. Self- (S) evolved independently in several of these adaptive branches.

~ f r e r Bmh, 19S5; after Grant and Grant, 1965.

a food source, and plants benefit by having their distribu- tions extended (because ants invariably drop some seeds on their way back to the colony). There are two major variations in this phenomenon: some ants feed upon the seeds (and so plant dispersal depends upon the frequency of seeds dropped); and some ants feed upon an elaioso~ne (fat body) that is attached to a seed, leaving the seed to germinate (Handel and Beattie 1990). This latter relation- ship, termed ~nyrnzecocho~y (from the Greek for "ant," 111Jrlnex, and for "dispersal," kore) insures complete dis- persal of all the seeds of a plant that u e harvested by ants. The ants benefit from consumption of the elaiosome, and the plants benefit from effective seed dispersal. Elaio- somes have been found in taxonomically diverse plants, including nonhern temperate wild flowers (e.g., Trilli~un and Viola), tropical epiphytes and vines, and Australizn arid shrubs (e.g., Acacia). The morphologically and taxo- nomically diverse origins of elaiosomes indicates conver- gent evolution: originally they protected seeds from predz- tion but gradually changed into food lures for ants. Even the biochemical attractants contained within the elaio- somes have been identified, with some proving to be the same even in plants found on different continents!

Other ants have special relationships with plants whereby they cultivate fungal gardens (providing nutrients for the ants) in exchange for shelter 2nd a supply of plant

Guinea, Huxley (1978) studied ant-plants that have evolved bu!bous chambers to house their ant partners, and Thompson (1981) explores how these associations may have arisen. Huxley and Cutler (1991) provide a review of a~t-plant interactivities.

Plants and Pollinators Rowers and their pol!inators (figure !0.15) have devel- oped specialized relationships over evolutionary time. The first winged insects were present 300 million years ago in the Carboniferous, long before the development of flowers (Smart and Hughes 1972). The first flowers appeared in the fossil record approximately 225 million years ago and were small (those of the gymnosperm order Bennettitales), so presumably they had insect pollinators. They are termed ei~~o~noplzilous (pollinated by insects that feed on pollen, nectar, or on the flower itself). At about the same time, the fust holometabolous insects appeared. The adults of these insects u1ere adapted to eat different foods than the larvae ste. This alteration of food supply between different life stzges greatly enhanced the opportunities of insects to spe- cizlize upon leaves as larvae and upon pollen as adults. The radiation of Lepidoptera, an insect order whose coevo- lution in pollination ecology is well studied, began during the early Cenozoic. Today we appreciate a wide variety of

Sexual reproduction in flo\vering plants is accomplished by the transfer of pollen from h e anher of a male flower to the stigma of a female flower. \%%en a pollen ga in contacts a stigma, the male germ cell in the pollen grain eventually unites wih h e female g e m cell or egg, and a fertile seed de- velops. \%%en the seed is exposed to h e proper conditions in or on an appropriate substrate, it \\.ill $ve rise to a new plant. The transfer of pollen from a male to a female flower is zc- complished primarily by h e \vind or by the activities of in- sects hat associate \vih plants. Esmples of \i.ind-pollinated plants include cereal plants such as \\heat and corn, ragweed (of hay fever fame), and many species of trees. The flo\vers of these plants, which are generally small \vith \veakly de\.el- oped petals, do not produce nectar, and produce dry pollen

. - \\.as studied by Heinrich (1979), who found several inter- esting patterns: flowers may evolve to allow visits only by pollinators, excluding "robbers" that take the rewards \vithout transferring pollen; plants evolve different levels of nutritional reward and densities of stands to regulate the number of flowers that insects are encouraged to visit; plants can control the amount of the calorific reward in re- lation to the s ize and behavior of the pollinator (e.g., a bumble-bee requires more energy ban a small fly).

Dis t inc t co r re l a t ions c a n be made between t h e anatomical and physiological characteristics of the flowers of a given species and the anatomy, physiology, and be- havior of their insect pollinators. Among the characteris- tics of flowers that attract insects are the follo\ving:

grains ha t are easily picked up by h e wind. 1. The production of particular scents The evolution of pollination involved the processes of 2. Color, size, and shape of petals

diverge~~ce (whereby organisms become different) and 3. Patterns of stn'pes or spots on petals collvergence (whereby organisms become more similar

1. Separation or nonscparation of petals o\*er time). Fossil e\.idence indicates that animal mutual-

5. Shape of flower ists changed more significantly over evolutionary time ns - -

compared to flo\vering plants, u.hich have remained re- markably stable o\.er the last f?\;. million years. However. some plant families exhibit classic specialization of repro- ducti\.e parts that attract specific pollinators. X good ex- ample is the phlox family, which evolved from one bee- pollinated flower to a variety of flower heads that ha\.? different types of pollinators (figure 10.15).

Many factors affect the complex relationships among flowers and heir insect pollinators. These include the nu&- tional rewards of pollen and nectar, as \\jell as the physicd features such as shape, color, and scent of flo\rlers; the phc- nology (i.e., seasonal availability) of flowers in relation to population dynamics of insects; the growth structure of flou- ers (clusters versus solitary); h e plant community surround- ing a flower (i.e., competition from other plants present); and, of course, competition from other pollinators. With 211 these factors to account for, it is obvious how complex p la t - pollinator relationships are for biologirs to unravel!

In pollination, both partners enjoy a mutual benefit from the relationship. The plant is propagated, and the pol- linator enjoys a calorific reward. Xn interesting aspect of the energetics of the plant-pollinztor mutualism \vas invcs- tigated by Baker and Baker (1973). It has long been h o \ r n that nectar contains various sugars and is used as an en- ergy source for pollinators, but insects also require th: molecules \vith which to synthesize proteins (amino acids). u.hich presumably had to be found elsewhere (e.g., in Lep- idoptera, from leatfes eaten by larvae). The Bakers sur- veyed 266 species of flowering plants in California and found that nectar contained amino acids in many species. and that re1atii.e amounts of amino acids were highest in specialized lepidopteran-pollinated flowers.

Plants and pollinators not only have complex relztion- ships in tenils of energy exchange, but plant structures en- sure long-tern1 consequences such as regulation of the PO!-

Scent production is often important in attracting a polli- nator and in determining behavior. Flo\i'ers pollinated by flies, \vhich breed in dung and canion, may actually mimic the odor of duns or rotten meat, so "fly flowers" commonly ha\.e a disagreeable odor to humans. Conversely, "butterfly and moth flowerst usually have a sweet smell. Honeybees are znracted to blue, purple, and yellow flowers, but cannot "see" red flo\vers. Many insects are attracted to patterns of ultraviolet light called nectar guides, which are believed to aid in directing the pollinator to the source of nectar (see chapter 6). The shapes of nectaries are often suited to restrict undesirable species and to encourage h e pollinator species.

Over the course of evolution, many pollinators have adapted to using particular flowers. X good example of this is the orchid family, the majority of whose members have specific pollinators and structurally complex flo\vers to pol- linate. Some orchids produce scents that mimic the effects of sex pheromones and also stimulate the insects visually and tactilely. Another interesting family is the figs, with 600 species, each of which has a specialized wasp pollinator. Pollinators that visit only one taxon of plant are termed inoirolecric (poljlecric pollinators visit many different taxa). What is the evolutionary advantage of visiting only one taxon? In the case of fig \rfasps, they not only pollinate the fig flowers but also lay eggs in the floral receptacle, called a ~ c o n i u i i ~ . This swcture provides a safe home for the wasp larva, ulhich consumes the ovule in which i t was placed. Upon emergence, it can easily find a mate and fly to another ripe fruit, thereby completin,o its life cycle wirhin one tree canopy. In this sense, monolectic behavior offers a secure existence to the host-specific fig wasp.

Even the foraging dynamics of pollinators can be a complex, coevol\~ed process. For example, euglossine bees forage for food throughout kilometers of tropical forest and frequently return to the same plants (Janzen 1971b). This

lination process. The promotion of ourcrossi~~e (wherebv a trap-linino behavior is advantageous because outcrossino,

rimy tropical tree species. Only one of several flowers must p r o j u ~ e nectar each day, and floral morphology can become sptci&zed for pollinators that can learn their way around a forest 2nd that are usually also efficient (and do not waste pollen), Because smaller amounts of energy are required to ensure pollination, uopical trees can reproduce much earlier i n heir life span and also save energy for other more com- pziirive aspects of sunival (e.g.. getting light 2nd nutrients). moral visibility is also less important and can be effective even under the shaded, complex tropical canopy.

Ths most common insect pollinators are members of the orders Coleoptera, Lepidoptera, Diptera, and Hymenoptera. Their most common adaptations include elongated mouth- 2 s s th5t allow hem to get nectar from flowers with deep :~:ctzries, plumose (fatherlike) hairs on the body to which ?ollen clings, and various specialized pollen-collecting :nd/or umsponing structures such as -the corbiculae (pollen ~sskets; see figure 2.35a) on the hind tibiae of honeybees. i3e rels!ionships of plants and insect pollinators are an excit- np aspect of ecologic21 research, because coevolution is a I!nmic piocess. and there are so many groups of insects 2nd plmls yet to study. The topic of insect-plmt interactions s covered by Priceet al. (1991) and Abraharnson (1989).

nsects in Forests as an Example of :ommunity Aspects of Insects and Plants lsects may form a community within a plant, and simi- !rly insects and plants together comprise p u t of a larger ~mrnunity or ecosystem. Although insect epidemics can :cur anywhere (even in deserts), forests generally contain e most diverse and abundant communities of inverte- ates. This is because forests contain a wide diversity of snts (trees, shrubs, herbs, mosses, lichens, ferns, and gae, each with its own insect fauna). Forests are slow- )wing and older than many surrounding early succes- )rial ecosystems; thus, they offer a large selection of >d sources of different ages, palatabilities, and nutritive 2lities. The relatively long time scale involved in forest -cession allows for colonization and more complex lev- of interactions of organisms to become established. The study of insects in forests is difficult, but it is in-

uing because of the anticipated diversity to be found :e. Whereas biologists can count and measure the num- of mollusks on the intertidal rock platform (a relatively 1-dimensional habitat), it is nearly impossible to view invenebrates of a forest due to its structural density, ~plexity, and height. These logistic problems have led 1)' scientists to study insects in other hzbiwrs; hence, a :ity of literature exists on forest insects, except for out- .ks when numbers are easier to obtain. In recent years, biologists have become particularly

rested in insects of rain forests for two reasons. First, hzbitats are endangered by human activity, and the to study them before they are reduced to fra, ~ m e n t s is

(Wilson 1988). Despite the fact that they occupy less than 7% of terrestrial habitats, rain forests have a variety of planrs and a relatively homogeneous wet climate that has enabled the evolution of a wealth of species. Wilson (1987) found 43 species of ants in one leguminous uee in the tropical rain forests of Peru, equivalent to the ant fauna of the entire British Isles! And Eruin (1982) found 1,060 species of Coleoptera by fogging four rain forest tree canopies in Brazil. \{%at was more astounding than the di- versity, hou~ever. was the fact that 83% werc erlderrlic to only one site. Only 3.2% of the species were shared among all four sites, indicating that almost every tree canopy in the tropics may have a specialized insect fauna!

In subtropical understory vegetation in Australia, Lolvman (1982) sampled insect densities that rmged from 65 to 650 per 100 m3 throughout different months of the yea . Thus, even in the relatively benign climates of rhe tropics and subtropics, insect seasonality is very pro- nounced. This suggests that biologists studying tropical in- sects need to consider many factors in their sampling de- sign and replication: numbers and diversity of tree species, age of forest. season of year, day versus night, phcnology of foliage, and weather, to name a few.

Herbivory (feeding damage) levels are often more com- monly measured in vegetation than are the insects that cause the damaze. This is panly because herbivores are often hard to see, feeding only inrermjttently or at night. Herbivory semes as ul important indicator of insect abundance, partic-

- ulvly in ecosystems such as rain foresrs, where the herbi- vores are especially difficult to observe and count. For ex- ample, high levels of herbivory in the cool temperate rain foresrs of Australia led Lowman to look for the insect(s) re- sponsible for this defoliation. Like many aspects of field bi- ology, sezrching for an insect in a complex mosaic of green foliage is equivalent to finding a needle in a haystack. Fortu- itous obsen.ations led to the discovery of a new genus of chr).somelid beetle whose larvae are prolific only during a 10-day period in spring (October) every year. This herbi- vore. X'orocas~ria ~lorlrofagi, eats only the Antarctic beech trees that dominate the cool temperate rain forests in New South \Vzles. The beetle lvvae emeree synchronously with the l e d emergence and defoliate over one-half of the new leaves, contributing significantly to the overall 30.5% an- nual l e d surface removal for beech canopies (figure 10.16).

Tnjs new beetle is just one of a vast fauna of get-to-be- discovered insects in this rain forest type alone! KO one can predict ihe actual numbers of insects and planrs that are be- coming extinct as rain forests are continually cleared, but the urgent need to document, study, and presene the ecol- ogy 2nd biodiversity of what remains has become m impor- tant issue of international priority today (see chapter 15).

Ho9w many insects are found in a tree? This question appezrs simple and straightforward, but is actually very complicrted and not well-defined. Many logistic aspects (previously mentioned) challenge biologists who want to

mount (see chapter :reatest wealhof in

Figure 10.16 Herbivoq in canopies cf cool temperate rain forest, dominated by Antarctic beech (Norhofagus moorei). Each number - represents mean amount of leaf surface area removed annually by insects (%). DcmfTorn hl. Louman, unpblithed.

D.a. = Dicksonia anfarcfica C.V. = Cuftsia viburnea N.m. = Nothofagus moorei 0.5. = Ouinfinia sieben' D.s. = Doryphora sassafras C.q. = Coprosrna quadrifolia

C.S. = Callicorna serra fifolia

Cool temperate herbivory - 27%

previously mentioned, and the methods of counting insects day, and forest structure and composition for replication. in trees are not well established. Ecologists face challenges This method fails to collect insects in the bark and roots, in merely locating all insects in a tree, much less counting 2nd may underestimate numbers of large insects that may them! Some insect orders characteristically feed on leaf be more resistant to insecticide. surfaces (Coleoptera, Heniiptera, Phasmida, and Lepi- Arthropod diversity increases froni the temperate re- doptera larvae) and are best sampled \~.ith sweep nets or gions to the tropical latitudes, as does the relative abun- with beating trays. Other orders zre often involved with dance of ants (see Tobin 1995). The recent development of temporary feeding on trees (e.g., pollinators) and occupy methods to access tropical forest cznopies has led to a re- the air spaces between trees (Diptera, Lepidoptera, some ne\red appreciation of the importance of ants in forest Hymenoptera). Some insects are nocturnal and may be col- ecosystems. Ants are the dominant group in many regions lected with light traps. Insects associated with the soils and of tropical rain forest, both in terms of abundznce and also roots of trees are difficult to quantify (but see Lou.man et in biomass. Ants represent up to 50% of the total biomass ,

al. 1987, and Badejo and Vanstrzelen 1993). Pitfall traps of arthropods collected from forest canopies, as well as up 1

are useful for collecting ground insects, but they are not to 50% of individuals at some sites (see table 10.2). comprehensive in their ability to quantify the abundance of The paradox of insect outbreaks in forest trees is a f organisms. Insects in the wood 2nd root systems of trees vast and complex topic, sumniarized in an entire volume 'i usually go uncounted. There are no comprehensive meth- (Barbosa and Schultz 1987). Trees are sessile. long-lived, ods to measure insect abundance: the use of sweep nets is and extreniely reliable food sources for insects, panicu- perhaps the most \r.idely used technique. It is obvious t h ~ t larly in forests \!?here the amount of foliage ranges from $ the numbers of insects in trees are vastly underestimated. 1-9 tons per hectare per year (Van Cleve et al. 1983). .:{

Fogging is perhaps the most comprehensive method More than 120 lepidopteran species feed upon Quercus .$ .:- of documenting the numbers of insects on trees. It involves i.olur, yet very few reach outbreak levels (Feeny 1970). 2 the dispersal of an insecticide throughout the upper tree Insects are very fecund and are reno\rned for their ability 2 . . canopy, usually with a spraying device. T h e insects 2re to adapt to large, persistent food supplies (agricultural 2 collected on sheets at ground level ~ f t e r the insecticide has fields, trees), so it is surprisinz that outbreaks are not more ?$ been effective. The sampling technique requires windless common. Of the vast numbers of herbivores on western '$ ..s conditions with careful attention paid to season, time of American trees, only 31 species are considered pests (Fur- .--:

. . . - .

.. . . .

T a b l e 10.2 Relative Abundance of Ants in CanoPY Aithropcd ~ornmlulitiesa

Percentage of Percentage of

Source Research site. Forest type total biomass individuals

hforan and Southwood (1982) - ' ' South Africa St 0 3 5 0-58 hforan and Southwood (1982) Britain Te 0 4 1 0-< 1 Erwin (1983) Manaus, Brazil Tr - 43

Hijii (1983. 1989) Nagoya, Japan Teb 0 0 Adis et al. (1984) Manaus. Brazil Tr 26-47 43-53 Stork (1987. 1991) Borneo. Brunei Tr - 7-32

hfajer and Recher (1988) Western Austnlia StC - 26-48 Stork (1988) Seram, Indonesia Tr - 42 Watanabe and Ruaysoongnern (1989) Northeatern Thailand Dt 17-46* - hfajer (1990) Q'estern Australia Ste 6-12 - h1ajer (1990) . Western Australia Stf 86 - hlajer (1990) . . Ghma Tri 5-72 - Stork and Brcndell(l990) ' . ' :Sulauesi, Indonesia Tr " - 1-32 Tobin (1991) b l a u , Peru Tr - 70 Kitching et al. (1993) Queers!and, .Australia Tr - Up to 26' Harada and Adis (1991) Cend hmzzonia Tr - Up to $6' - J. E. Tobin (unpublished data) B m Colorado, P m m a I : 16-57 - J. E. Tobin et al. (unpublished data) Tarnbopam, Peru Tr - 42-94

8Summarj table of relati\c abundance of anu i r ~ canopy &o+ m p l u collccred unng ~=u; t ic~ l i l i - . ~ l d o u n techniquu. Sumben a:e t.4~ prcentage coniribut~on of anu to the total arrhropd biomass and to the to:il number of md:\iJuaI L-?ZJ$S 1-1 ramplu B ~ o n b u ir g i ~ e n ac dry ueip4r. except u noted. h4ort srudls are of 1oula;ld uopical -opiu, but wvcnl subt.o;ial and temperate forut r r ~ 2 1 u k-e ~~: l .~ded fo: conparson. T h e \I:IOLU r ) ~ oilouland rain fo-u: are not dutinguhhed Tr - lowland uopical rain for- Dr - d;) co?iul forest; St - s u b z q i d forG Te - temprate forur tThue wqsrudia wye conducted in plktationr ? f J a p a n y c y , ~ a n d J a p u e cedar, h . ;cc t~ \e l~ : &40ugh the acscciated canopy arthropd commurikies ue:e laze dl\ecrj;,'&u arire'a&ni :-5c Z:': --,-" '- ' 'Eucalpt forau at Dqandn Srate FoIesc, U'utern Australla. *Wet ucighr ': . . - _ _ -. I-

( h u l y p t forum ;c-~r)'andn scat; &rut and at ~am&!len, F7crren;~usml~a. 'Mango plantation. .. ' - .: . I _

r & a plantation; figures are for the dominant ant s p i a on'!. 'i- hlncluda data from a variety of agriculrud. suamp foruc, and louland and midele\arion fors, s!ru. 13 general, anu uerc l u s dominant at h u e ri tu than at orher Smtheasr'Asian and Neouopical s i t s ; Dipten uac the m m a 5 ~ ~ d a n r group. '.4u~4orr do nor give minimum figures. . From J E Tobln, 'A Neouopical Rainform Canopy Ant G x ~ u n i r ) : &me Esolqiul G-sr?eratlo~' ~n C R. Huxley and D. F. Cuder. Ed?. Am-Plmu 1nsc;ew.s. Gp)nght 0 1991 O~ford University Pras, Oxford, E n g h d Reprinted by pemision.

ss and Carolin 1977). Only 16 of these are defoliators 4 th 15 as feeders in bark or wood), and 3 of the 16 are :reduced (e.g., gypsy moth). The number of native pests I this apparently abundant food supply is exuemely low!

Whether or not insects are considered pests is an ail-

~opocerrrr-ic (human-centered) view, and fluctuations in sect numbers are often a consequence of human activi- :s. The eucalypt dieback in'Australia and the gypsy moth Itbreaks in the eastern United States are good examples. 5 long as insects continue to fluctuate in numbers over ne and spacz, and as long as they continue to compete ith humans as herbivores, we \trill continue to study the lrnplex interactions of insects on plants.

lieferences

~rahamson, W. G., editor. Plant-animal interactions. New (ork: hlcGraw-Hill; 19S9.

Bruil. wit! critical considerations on the pyrethrum-fogging technique. Studies on Neotropical Fauna and Environment 19:223-236; 1984.

Allison, .A,. Samuelson, G. A.; hliller, S. E. Patterns of beetle s,rcies diversity in New Guinea rain forest as revealed by m o p y fogging: preliminary findings. Selbyana 14:1620; 1993.

Xnir,~!fishnan. T. N., editor. Biology of gall insects. hndon : !3u.~.d .Arnold; 1984.

Bath. C. E. Effects of plant density and diversity on the population 6!-r,irnics of a specizlist herbivore. the smped cucumber bcctle. . i ~ ~ / ~ l l : l l . J rirrara (Fab.) Ecology 61 :1515-1530; 1980.

Bxh . C. E. Host plznt growth form and diversity: effects on ibvndznce and feeding preference of a specialist herbivore, i.cc1~1n1rs i.irrara (Coleoptera: Chrysomelidae). Oecologia 50:37&375; 1981.

Bsdejo. 5f. A.; Vanstraalen, N. bl. Seasonal abundance of s?i%g~ils in two contrasting environments. Biotropica 25:222-228: 1993.

Bd:er. H. G.: Baker. I. Amino-acids in nectv and their e \o lut ior .q signiiicance. Nature 241:543-545; 1973.

B ~ x i j e e . B. .An analysis of the effect of latitude, age, and area on

- -. .. . --- jis. H.,.Lubin, Y. D.. h j o n t 8 g q y . G. G. Anhropods f r o ~ the . the numkr of pest species - - of tea. . -. - . . . . . -- - - _ _

, ~ a r b o s a , P.; Schultz. J. C., editors. Insect outbreaks. San Diego. Grant. V.; Grant, K. Flower pollination in the Phlox family. CA: Academic Press; 1987. Sew York: Columbia University Press; 1965.

~ b , F. Insects and flowers. Princeton, XJ: Princeton Unjversiry H&ston, N. G.; Smith, F. E.; Slobodkjn, L. B. Community Press; 19S5. scrucru're, population control, and competition. American

Basset. Y. A composite interception w p for sampling arthropods Saturalist 94t421-425; 1960. in tree canopies. Joumal of Australian Entomological Socier). Hmdel, S. N.; Beanie. A. J. Seed dispersal by anu. Scientific 27:213-219; 1988. r American 263:76-83; 1990.

Bazzaz, F. A. The response of natural ecosystems to the rising Hxada, A. Y.; Adis, J. The ant fauna of tree canopies in central global CO? le\.els. Annual Review of Ecology and Systematics mazonia: a first assessment. Abstracts International Congress 21:167-196; 1990.' of Ecology (A'TECOL), 6th, hlanchester, England; 1991.

Bernays, E. A. The insect on the plant-a closer look. Huborne, J. B. Introduction to ecological biocheritistry. 3d ed. Proceedin_~s of the 5th International Symposium on Insect-Plmt London; San Diego, CA: Academic Press; 1988. Relationships, Wageningen, The Netherlands. Wageningen, Haulcioja, H. Induction of defenses in trees. Annual Review of Holland: Centre Agriculrural Publication Documentation; 1982. Entomology 36:25-42; 1991.

Bernays, E. X.: Chapman, R. F. Host-plant selection by Hcattvole, H.; Lowman, hl. D. Dieback4eath of an Australian phytophapous insects. Xew York: Chapman B: Hall; 1991. Izndscape. Sydney. Australia: Reed Publishing; 1986.

Borror, D. J.; DeLong, D. bl.; Triplehorn, C. A. An introduction Heanvole. H.; Lowman, M. D.; Donovan, C.; hlcCoy, hl. 1997. to the srud!. of insects. 4th edition. Xew York: Holt. Rineha,, Phenology of leaf flushing and macroanhropod abundances in and \'.'inston; 1976. cmopies of saplings in a Eucalypfus woodland and in the open.

Clements, F. Plant succession: an znalysis of the development of Selbyana 18(10). (In Press.) vegetation. Carnegic Institution of Washington Publication Heinrich, B. Bumblebee economics. Cambridge. SIX: H m a r d KO. 242. 1916.512 pp. V~ver s i ty Press; 1979.

Coley, P. D. Herbivory and defensive chvacteristics of tree Hcspenheide, H. A. Bionomics of leaf-mining insects. Annual species in a lo\\lland uo;ical forest. Ecological hlonographs Review of En~omology 36:535-560; 1991. 53:209-233; 1983. Hijii. N. Arboreal arthropod fauna in a forest. I. Preliminary

Edwards, J. S. .ahropods of alpine aeolian ecosystems. Annod --.,, . - :L.\.s:ions on seasonal fluctuations in density, biomass, and : Review ofEntomology 32:163-169; 1987. faunal composition in a Chatnaeqpans obrusa plantarion. Erwin, T. L. Tropical forests: their richness in Coleoptera and Japanese Journal of Ecology 3 3 : 4 3 5 4 ; 1983.

other arthropod species. Coleopterisrs Bulletin 36:74-75; 1961. Eijii, N. Arthropod communities in a Japanese cedar Erwii, T. L. Beetles and other insects of tropical forest canopies et (Crypfotneia japonica d. Don) abundance, biomass,

hlanaus, Bnzil, sampled by insecticidzl fogging. In: Sunon, S. L.; 27d some properties. Ecological Research 4:243-260; 1989. IVhimore, T. C.: Chadwick A. C., editors. Tropical rain forest: Hodson, A. C. Minnesota springs: a fifty-year record. Minnesota ecology aid management. Oxford: Black-,ell; 1983:p. 59-75. Honiculturist 119:15; 1991.

Erwin. T. L. Measuring arthropod biodiversity in the tropical Hoot~er. C. hl.; Crossley, D. A., Jr. Leaf litter decomposition and forest canopy. In: Forest canopies. Lowman, M. D.; Nadkami! rnicroanhropod abundance along an altinidinal gradient. In:

: N.. editors. San Diego: Academic Press; 1995:p.,109-127. Significance and regulation of biodiversity. Collins. H. P.; _..&

'Feeny, P. Sevonal changes in oak leaf tannins and nutrients u 2 ..y

Robertson, J. P.; Klug, hl. J., editors. Kluwer Academic .+ cause of spring feeding by winter moth caterpillars. Ecology Publications: Netherlands. 1995: p. 287-292. ;< . B 5 1565-58 1: 1970. Howe, H. F.; Westley, L. C. Ecological relationships of plants 23 . .

Fox, L. R.; I.lacauley. B. J. Insect grazing on Eucol~pfus in md animals. New York: Oxford University Press: 1988. .y response to variation in leaf tannins and nitrogen. Oecologia Huxley. C. R. The ant-plants Myrmecodia and Hydnophytum 7~2 ..-,

. - 29: 145-162; 1977. (Rubiaceae) and the relationships between their morphology.

Fox. L. R.; hiorrow, P. A. Esrimates of ,prrzing damage to mt occupants, physiology, and ecology. New Phytologist Euca4prus trees. Australian Journal of Ecology 8:139-144; 19E5. 80:231-268; 1978.

Fox, L. R.: Irlorrow. P. A. On compYing herbivore damage in Hcxley, C. R.; Cutler. D. F.. editors. Ant-plant interactions. Australian and north temperate systems. Australian Journal of Sew York: Oxford University Press; 1991. Ecology 11:387-393; 19S6. Jwzen. D. H. Coe\+olution of mutualism betbveen ants and

Fraenkel, G. S. The iaisor~ d'efre of secondvy plant substances. zcacias in Central America. Evolution 20:249-275; 1966. Science 129:1466-1470; 1959. ;~ .~zen . D. H. Seed predation by animals. Annual Review of ,.:

Furniss, R. L.; Carolin, V. I f . Western forest insects. Ecology and Systematics 2:465492: 1971a. hliscell~neous Publications-United States Department of Agriculture 1339; 1977. topical plants. Science 171:203-205; 1971b.

Gaston. K. J. The magnitude of global insect species. Johson, R. J.; Lincoln, D. F. Sagebmsh carbon allocation panerns

roil mineral limitation. Oecologia 87:127-131; 1991. - aderlopoda \\.on the selectional race svith heliconiine

butterflies? Science 172:585-586; 197 1. Gleason, H. A. The individualistic concept of the plant association.

Bulletin of the Torrey Botanical Club 53:331-368; 1926. Graebe. J. E. Gibberellin biosynthesis and control. Annual

Review of Physiology 38:119-465; 1987.

cr.idar~ia larvae (Lepidoptera: Noctuidae). Journal of Insect physiology 35599-708; 1989. .itching. R. G. L.; Bergelson. J. hl.; Lowman, hf. D.; hlcIntye. S.; C m r h e n , G. The biodi~~ersity of anhropods from Ausnalian r:inforest canopies: general introduction, merhods, sites and

results. Australian Journal of Ecology 1S:ISl-91; 1993. :nlger. F. J.; hlitchell, D. T.; J m i s , J. U. hf., editors. >1editerranean-type ecosystems. Berlin: Springer-\'erlag; 1983. afont, R.; Horn, D. H. S. Phytoecdysteroids: structure and occurrence. In: Koolman, J.. editor. Ecdysone: from chemistry 10 mode of action. Stuttgart: Thieme-Verlag; 1989:~. 3 9 6 1 .

Aitron. J. H. Plant architecture and the diveniry of phytophagous insecrs. Annual Review of Entomology 28323-29; 1983.

,tiin, D. A. The role of trichomes in plant defense. Quarterly Review of Biology 48:3-15; 1973.

-e\vis. T. Thn'ps: their biology, ecolo_ey, and economic imporiance. London: Academic Press: 1973.

_iss. \V. J.: Gut, L. J.; \\'estigard. P. H.; Warren, C. E. Perspectives on anhropod community structure, organization, and de~reloprnent in agricultural crops. Annual Review of Entomology 3 1:155178; 1986.

-cagino, J. Species richness of ants (Fomlicidae) in a tropical forest canopy, using insecticidal fogging and richness ctrimation procedures. Selbyana l5(2):A13: 1991.

Loivna , hl. D. Seasonal variation in insect abundance among several Ausualim rain forests, Ivith panicular reference to ph!.tophagous types. Ausualim Journal of Ecology 7:353-361; 19S2.

-o\vman, hl. D. Spatial and temporal van'abili~y in herbivory of five canopy eees in Ausualia. Australian Joumzl of Ecology 10:7-24; 1985.

-o\vman, M. D.; Burgess, A. D.; Hiegins, W. D. The biomass of Yet\. England peppermint (Eucalyprus 11or.a-anglica) and associated insect damage related to rural diebzck. Australian Joumal of Ecology 12:361-373; 1987.

-owman, hl. D; Fox J. Variation in leaf toughness and phenolic content among 5 species of Australian rain forest uees. Australian Journal of Ecology 8:17-25; 1983.

-ownan, M. D.; Heatwole, H. Spatial and temporal variability in defoliation of Australian eucalypts. Ecology 73:129-142; 1992.

iowrr.ln, hl. D.; hfoffen, M. The ecology of tropical rain forest canopies. Trends in Ecology Evolution 8:103-10s; 1993.

Lowman, M. D.; Yadkami, N.. editors. Forest canopies. San Diego: Academic Press: California. 1995.

ilajer, J. D. The abundance and diversity of arboreal ants in nonhem Ausualia. Biouopica 22:191-199; 1990.

iiajer, J. D.; Recher, H. F. Invenebrate communities on western husualian eucalypts: a comparison of branch clipping and chemical knockdotvn procedures. Ausualian Journal of Ecology 13:269-278; 1988.

~ ~ ~ l : s o n , \V. J. Herbivory in relation to plant nitrogen content. Annual Review of Ecology and Systematics 1 l:l19-161; 1980.

i l c ~ e i l l , S.; Southwood. T. R. E. The role of nitrogen in the development of insecdplant relationships. In: Huborne, J. B.,

Biochemical aspects of plant and animal ~oe\~olution. London: Academic Press; 1978:~. 77-98.

hfelcalf, Z. P. 1910. How many insects are there in the world? Entomological News 5 1 2 19-222; 1910.

h7e!'er, J. Plant galls and gall inducers. Berlin: Gebruder Bomtraeger; 1987.

hliner. C. B.; Farrell, B.; \Viegmann. B. The phylogenetic study of adaptive zones: has phytophagy promoted insect diversification? American Naturalist 132:107-128: 1988.

hlitton, J. B.; Sturgeon. K. B. Bark beetles in North American conifers: a system for the study of evolutionary biology. Austin, TX: University of Texas Press; 1982.

hloran, N. X.; \%itham, T. G. Population fluctuations in complex life cycles: an esample from Per~~pltigus aphids. Eco lo~y 69:1214-1218; 1988.

hloran, V. C.; South\vood, T. R. E. The p i l d composition of arthropod communities in trees. Journal Animal Ecology 51:289-306; 1932.

hlorro\\*, P. A. The role of sclerophyllous I e a ~ e s in determining insect grazing damage. Ecological Studies 43:509-511; 1983.

hlorro\\-. P. X.; Fox. L. R. Effects of variation in Eucaljprus essential oil yield on insect growth and herbivore load. Oecologia 45: 109-129; 1980.

biorro\\~, P. A.; Fox, L. R. Estimates of pre-settlement insect damage in Ausualian and Nonh American forests. Ecology 70: 1055-1060; 1989.

Pate. J. S. Uptake, assimilation, and transport of nitrogen compocds by planls. Soil Biology and Biochemistry 5:109-119: 1973.

Price, P. \V. Colonization of crops by anhropods: non- equi1ib;ium communities in soybean fields. Environmental Entomo!ogy 5:605-611; 1976.

Price, P. I{'. General concepts on the evolutionary biology of puasites; Evolution 31:405-420; 1977.

Price, P. W. Evolutionary biology of parasites. kfonographs in Population Biology 15. Princeton, NJ: Princeton University Press; 19EO.

Price, P. \V. The role of natural enemies in insect populations. In: Baibosa, P.; Schultz, J. C., editors. Insect outbreaks. San Diego. CA: Academic Press; 1987:p. 287-312.

Price. P. W.; Cobb, N.; Craig. T. P.; Fernandes, G. \V.;lLami, J. K.; hlopper, S.; Preszler. R. W. Insect herbivore population dynamics on trees and shrubs: new approaches relevant io latent r?d eruptive species and life table development. In: Bemays, E. A.. editor. Insect-plant interactions. Boca Raton, FL: CRC Press; 1990:p. 1-38, vol. 2.

Price, P. W.; Lewinsohn. T. M.; Fernandes. G. W.; Benson. I\'. W.. editors. Plant-animal interactions-evolutionvy ecolo:y ia tropical and temperate regions. Somerset, NJ: John ISiley &: Sons; 1991.

Richards. 0. \V.; Davies, R. G. Imm's general textbook of entonolopy. London: Chapman and Hall; 1977, vo1.-3.

Riemer, J.: \i:hinaker, J. B. Air pollution and insect herbivores: obsened interactions and possible mechanisms. In: Bemays, E. A., editor. Insect-plant interactions. Boca Raton, FL: CRC Press; ISS9:p. 73-106, ~ o l . 1.

Roor. R. 8. Organization of a plant-arthropod association in simple a d diverse habitats: the fauna of collards (Brassica olerocea). Ecological hfonographs 43:95-124; 1973.

Rcseiiih~l, G.A.; Berenbaum, hl.R.; editors. Herbivores, their ir,;e;ac:icn with secondary plant metabolites. 2nd edition. Szn Diego: Academic Press, 1991.

Ryan, C. X. Proteinase inhibitors. In: Rosenthal. G. A.: Janzen. D. H., editors. Herbivores: their interaction with secondary plant 2e:abolites. New York: Academic Press; 1979: p. 599-61 8.

Ryan, C. A. Insect-induced chemical signals regulating natural plant protection responses. In: Denno. R. F.; hfcClure. 51.. editors. Variable plants and herbi\,ores in natural and managed systems. Xew York: Academic Press; 1983:p. 43-60.

Schulrz, J. C. Impact of variabk plant defensive chemisVq on susceptibility of insects to natural enemies. ACS Symposium Series 20837-51; 19S3.

Scriber, J. hl.; Slansky,F., Jr. The nutritional ecology of immature insects. .knual Revie\\, of Entomology 26:183-212; 1981.

Singh, B. B.: Hadley, H. H.; Bernard, R. L. 5lorpholog!' of pubescence in soy beans and its relationship to plant vigour. Crop Science 11:13-16: 1971.

Sman, J.; Hughes, N. F. The insect and the plant: progressive paleoecological integration. In: v2n Emden, H. F., editor Insecl/plant relationships. Symposia of the Royal Entomological Society of London 6: 1972:p. 112-155.

Southwood, T. R. E. Bionomic strategies and population parameters. In May, R. M.. ed. Theoretic31 Ecolo_ey. 2nd ed., Oxford: Blackwell; 19Sl:p. 20-52.

South\vood. T. R. E. The number of species of insect associated with various trees. Journal of Animal Ecology 30:l-8: 1961.

South\vood, T. R. E. The insecrlplant rslationship-an e\~olutionary perspective. In: van Emden. H. F.. editor. S!.rr,poria of the Royal Entomological Sociev of London 6: 1973:~. 3-50.

Sourh\vood. T. R. E.; Kennedy, C. E. J. Trees as islands. Oikos 41:359-371; 1983.

South\vmi,T. R. E.; h lom, V. C.; Ken~edy, C. E. J.The assessrnenrof arboreal insect fauna: compaisons of knockdo\\n sampling and faunal lists. Ecological Entomology 7333 1 4 . 1983-

Southuvood, T. R. E.; bioran V. C.: Kennedy, C. E. J. The richness, abundance. and biomass of rhe arrhropod comrnuniti:~ on trees. Journal of Animal Ecology 51:635449; 1981b:

Stork, X. E. Guild structure of anhropods from Bornean riin forest uees. Ecological Entomology 12:69-80; 1987.

Stork, N. E. Insect diversity: facts, fiction and speculation. Biological Journal of the Linnean Society 35:321-337; 1988.

Stork, N. E. The composition of the anhropod fauna of Bornean lowland rain forest trees. Journal of Tropical Ecology 7, 161-180; 1991.

Stork, N. E.; Brendell, hi. J. D. "Variation in the insect fauna of Sulawesi trees with season, altitude and forest type." In: Knight, W. J.; Holloway, J. D., editors. Insects and the rain forests of South East Asia (\\'allacea). London: Royal Entomological Society of London: 1990:p. 173-1 90.

Strong, D.; Lauton. J. H.; Southwood. R. Insects on plants. Oxford: Blackwell Scientific: 1981.

Tallamy. D. W.; Raupp. 31. J., editors. Phytochemical induction by herbivores. Somerset, NJ.: John IYiIsy & Sons. Inc.; 1991.

Thompson. J. N. Reversed animal-plant interactions: the evolution of insectivorous and ant-fed plants. Biological Journal of the Linnean Society 16:117-155: 19S1.

Thompson, J. N. Interaction and ~oe\~olution. S e w York: John ll'iley 6 Sons; 1982.

Thompson, J. N. Concepts of coevolution. Trends in Ecology and Evolurion 61:179-183; 1989.

Tilman, G. D. Cherries, ants, and tent ca!c~ill3rs: tinu'ng of nectar production in relation to susceptibility of czterpillars to ant predation. Ecology 59:686-692; 1978.

Tobin. J. E. A neotropical rainforest canopy ant community: some ecological consideration. In: Huxley, C. R.; Cutler, D. F., editors. Ant-plant interactions. Oxford: Oxford University Press: 1991:p. 536538.

Tobin, J. E. Ecology and diversity of tropical forest canopy ants. In: Lowman. 51. D.; Nadkami, N., editors. Forest canopies. San Diego: Academic Press; 1995:p. 129-147.

Turner. J. P.. G. Adaptation and e\,olution in Heliconius: a defense of neo-Dm\.inism. Annual Review of Ecology and Systematics. 1299-121; 1981.

V3n Cleve, K.: Oli\,er, L.: Schlentnez, R.; Viereck, L. A.; Dlmess, C. T. Pioduction 2nd nutrient cycling in uiga forest ecosyaems. Canadian Journal of Forest Research 13:717-766; 1983.

l\7ainhouse. D.; Cross, D. J.; Howell, R. S. The role of lignin as a defense against the spruce bark beetle Derrdrocrorrus n~icons: effect on larvae and adults. Oecologia 85:257-265; 1990.

\Yaldbauer, G. P. The consumption and utilization of food by insects. Advances in Insect Physiology 5229-289; 1968.

Watanabe. H.: Ruaysoongnem, S. Estimation of arboreal vlhropod density in a dry cvergreen forest in nonheastem Thailand. Journal of Tropic4 Ecology 5:151-158: 1SE9.

\\'aterhouse, D. F. The insects of Australia. 31slbourne, '

.Austr~!ia: Uni\.ersity Press; 1970. i\-atem~:n. P. G.: liole, S. Extrinsic factors influencing production

of secondary metzbolites in plmts. In: Bernays. E. A., editor. Insect-plant inieractions. Boca Raton, FL: CRC Press; 1989: p. 107-131. vol. 1.

\yedin. D.: Tilman, G. D. Niuogen cycling, plant competition, and the stability of d l g a s s prairie. In: Smirh. D. D.; Jacobs, C. A., editors. Proceedings of the Tuelfrh Honh . a e r i c m Prairie Conference: Recapmring a Vvlishing Herirage. Cedu Falls. LA: University of Sonhern Iowa Press; 1992.

\'vreins, J. A.; Cates, R. G.; Rotenbeny. J. T.; Cobb. N.; Van Home, B.: Red*. R. A. Anhropod dynamics on sagebrush (Arler~~isio ~riderl~ara): effects of plant chpisr ry and avian predation. Ecological hlonogaphs 6:299-321; 1991.

Weis, A. E.; Berenbaum, hl. R. Herbi\~orous insects and green plants. In: Abrahamson, W. G., editor. Plant-animal interactions. New York: McGraw-Hill; 19S9:p. 123-162.

\\hitham, T. G. Habitat selection by Pemphigus aphids in response to resource limitation and cnmpetition. Ecology 59:11611176; 1978.

Vhitham. T. G. Host manipulation of parasites: within-plant variarion as a defense against rapidly evolving pests. In: Denno, R. F.; I\lcClure. hf. S., editors. Variable plants and htrbi\,ores in natural and managed systems. New York: Academic Press; 1983:~. 3 5-39.

r\'k,itham, T. G. Plant hybrid zones as sinks for pets. Science 74:1?9-353; 1989.

Kilson. E. 0 . The arboreal ant fzuna of Peruvian Amazon foiests: a fi:sr assessment. Biotropica 2:745-25 1; 1987.

r\'ilson, E. 0 . . editor. Biodib~ersiry. Washington, DC: National .icademy of Science Press; 198s.

L'ilson, E. 0. The diversity of life. Cambridge, 31A: H m a r d Uni\.ersiiy Press: 1997.

Z:iolfer. H. E\,olutionary and ecological relationships of h e insect f2una of thistles. Annual Review of Entornology 23:103-122; 19SS.