Insect–plant interactions can range from associations that are highly detrimental to herbivores to those in which herbivores overexploit their plant hosts. In this chapter, the Committee on California Agricultural Research Priorities: Pierces’ Disease explores the naturally occurring mechanisms that control the outcome of insect–plant interactions that could be adapted to the management of the damage that attends pest–crop interactions. More specifically, the objective is to explore ways in which the links between the glassy-winged sharpshooter (GWSS), Homalodisca coagulata (Say), and its hosts, particularly its economically important crop hosts, can be eliminated, inhibited, or minimized; that is, to develop management strategies that reduce or impede this vector of the Xylella fastidiosa (Xf) bacterium. If management strategies are to be effective in sustainable pest management schemes they should include a cohesive, complementary, and coordinated set of tactics. Although discussions are provided on a variety of tactics, it is assumed that the overall approach to the resolution of the problem will involve implementation of an ecologically based pest management program.
Ecologically based pest management (EBPM) is “the release of living predatory, parasitic, pathogenic, or antagonistic organisms (biological control organisms), the deployment of biologically derived products, such as toxins or semiochemicals (biological control products), and the planting of resistant crop varieties (resistant plants)” (NRC, 1996). Thus, EBPM includes approaches that go beyond traditional biological control––the intentional use or manipulation of predators, parasitoids, pathogens, antagonists, or competitors (Van Driesche and Bellows 1996)––to control insects, weeds, or diseases. EBPM has been defined or described by a variety of names including biointensive integrated pest
management (IPM) and biorational IPM. EBPM is an IPM scheme that considers all possible management options and tactics that are complementary, effective, and sustainable with regard to the management of a pest and the quality of the environment. To be effective, EBPM tactics or agents might require environmental conditions that do not exist in agroecosystems. Reestablishing those conditions could require specific inputs or modifications of management tactics. EBPM typically requires more intensive management and planning and a fundamental knowledge of interactions among crops, pests, natural enemies, and the environment—requiring more time and record keeping. Some EBPM practices (such as augmentation biological control) could require an infrastructure that facilitates rearing, storage, and distribution of natural enemies (Neuenschwander et al., 1989).
Unlike strategies that rely solely on the use of insecticides, most EBPM is either specific to a pest or nontoxic, so it is less likely to damage the environment. Although many natural enemies can be susceptible to insecticides used against target pests, the judicious use of selective or specific types of insecticides typically has been considered an important component of integrated or EBPM schemes. Because most EBPM practices do not result in toxicity or pathogenicity to mammals, particularly in comparison to the use of conventional pesticides, they are more acceptable in terms of public health and environmental safety. Furthermore, insect pests are less likely to develop resistance to biological agents or to ecologically based controls. If resistance does develop, it is likely to occur significantly more slowly than would be occasioned by the conventional use of pesticides. Nevertheless, EBPM could pose risks, as do most environmental manipulations.
TACTICS AND STRATEGIES FOR ECOLOGICALLY BASED PEST MANAGEMENT
Ecologically based pest management can involve the use of living natural enemies (biological-control organisms), resistant crop varieties, or any product derived from those and other organisms. The tactics are deployed in any fashion that enhances their effectiveness and sustainable use against pest species. Those tactics and management approaches include the following:
host plant resistance (through conventional breeding or genetic modification)
vegetation management (in- or off-crop field refuges, trap crops, alteration of refuge vegetation, plant barriers)
traditional biological control (classical, augmentation, conservation)
sterile male technique
mass trapping using a compound derived from a living organism
direct use of chemical agents (synthetic toxins, biorational insecticides, behavior-modifying chemicals, insect growth regulators)
pheromone use and mating disruption
Although not all of those tactics are ready for implementation in the control of GWSS and Pierces disease, the tactics that could be used are reviewed below.
HOST PLANT RESISTANCE
Host plant resistance (HPR) is an important EBPM tool for minimizing crop damage caused by insect pests and diseases because it makes plants less suitable for or more tolerant of the pest. HPR has a long history of success in grape crops for management of phylloxera, grape leafhoppers, and various diseases (e.g., Granett et al., 1996; Jermini et al., 1996, Martinson and Dennehy, 1995; Walker, et al., 1994). HPR can be developed in three ways. Biosassays or direct observation in the field of apparent resistance or tolerance to a pest can be used to identify plant material for propagation. This has been the main approach to developing pest-resistant grapes because vegetative propagation is the norm, sexual reproduction delays the opportunity to perform crosses, and the value of the crop is directly related to its genetic pedigree (interspecific hybrids are not favored). Collecting and propagating resistant material can provide the quickest route to HPR, but the bases of resistance might not be known and resistance breakdown is common. HPR also can be developed from resistant or partially resistant plant material through standard breeding (crosses) if the plant’s biology is compatible (early sexual reproduction, outcrossing, compatible genotypes, inheritance relatively simple). The third way to develop HPR is to modify plant traits using direct molecular manipulation. Genetically transformed or genetically modified (GM) plants express naturally occurring resistance traits, but to different degrees or in different combinations than found normally, or they might express entirely novel traits (the protein toxins of Bacillus thuringiensis are an example). There is considerable debate about the risks, costs, and benefits of such approaches, especially for application in food crops. Resistance phenotypes can sometimes be manipulated through cultivation practices, such as fertilization.
No matter how HPR to a given pest is developed, it is subject to periodic, repeated breakdown as a result of pest evolution or the introduction of new pest species or genotypes (e.g., Granett et al., 1996). HPR as a management tactic requires constant evaluation and modification for sustainability. Although constant searches for new plant material or germplasm could remain part of this process, common sense suggests that understanding the plant traits that do or can confer resistance would greatly accelerate HPR development. Successful HPR use in EBPM requires a detailed understanding of how the pest exploits the plant and how plant traits interfere with pests.
Plant-Herbivore Interactions and Host Plant Resistance
Insects have been consuming plants for 500 million years, and during that time diverse interactions between the two groups have evolved. As far as the committee knows, all plant species and all possible plant tissues are exploited as food, shelter, or reproductive sites by insects. Insect diets range from extremely narrow (limited to a particular tissue on a particular plant species) to very broad (feeding successfully on many plant species from several plant families). A striking feature of insects as herbivores is that relatively narrow diets (a single or a few plant species) are more common than are broader diets. Diet breadth is evolutionarily fixed; insects usually cannot broaden or narrow their diets during a lifetime, and often require significant genetic change (evolution) to switch or add host species to their diets. Identifying the proximate and ultimate factors that determine the structure of insect diets is an active area of in plant-herbivore research, and it is safe to say that a complex picture is emerging. Many factors can act as barriers to plant exploitation and thus are thought to influence the evolution and expression of insect host range.
Plant chemistry is a crucial barrier to plant use and is thought to be an important influence on insect's host ranges. Plants are generally suboptimal food either because they contain low concentrations of critical nutrients such as nitrogen or protein or because they produce chemicals (allelochemicals or secondary metabolites) that deter feeding, inhibit digestion, or are toxic to the insect. Insects have developed specific adaptations that permit specialization on hosts that have particular chemistries, Those adaptations include the ability to avoid toxic tissues and a variety of biochemical traits that counter the negative effects of plant chemistry.
Plant quality also can have indirect effects on herbivorous insects. Insect susceptibility to parasites, predators, disease, and insecticides has been shown to be influenced by the host plant (Wright and Verkerk, 1995). Nonlethal effects on fitness (those that reduce fecundity, for example) could influence insect population growth. It is likely that most aspects of insect performance are influenced by diet quality and that plant chemistry is a central factor in limiting consumption by insects.
Although research on the effects of plant chemistry on insect herbivores originally focused on constitutive traits (those that are present constantly and encountered by an insect when it first goes to the plant), studies over the past two decades have shown that changes in plant chemistry that are induced by the feeding insects are at least as important (Karban and Baldwin, 1997). All plants so far studied alter the concentrations or composition of allelochemicals when attacked, sometimes in an insect-specific fashion. Those changes can reduce diet quality or enhance toxicity or deterrence in a matter of minutes or hours. Related changes are elicited in plants by pathogen attack, and there is evidence that the suitability of plants for insects could be influenced by plant responses to pathogenic and symbiotic microbes (Bostock, 1999).
Most plants also emit volatile substances that attract the enemies of insect herbivores (parasitoids, predators) and could cue defense responses in
nearby plants (Dicke and Bruin, 2001). Those indirect effects on insect herbivores could be more influential in protecting plants from insects than are the more direct constitutive or inducible defense traits (Karban and Baldwin, 1997). Volatile substances also could have a direct effect by repelling or poisoning sessile, sucking insects such as aphids (Birkett et al., 2000). Recent research indicates that plants also emit complex volatile mixtures when they are attacked by microbes (J. Huang et al., 2003); one study suggests that such pathogen-elicited emissions could have an effect on insect vectors of plant diseases (Eigenbrode et al., 2002).
Although the nutrient (protein, carbohydrate, fat, mineral) composition of plant tissues often is suboptimal for insect growth, it is assumed that all insects experience similar selection on their ability to extract sufficient nutrition from plants. Together with the need of plants to produce a minimal nutrient mix for their own functioning, the nutritional content of plants is rarely cited to explain evolution of insect host range. However, nutrition does shape insect adaptation and diet. It is true that even extremely nutrient-poor plant tissues are consumed by some insects, the insects that do so generally are limited to those tissues, presumably because adaptation comes at the cost of restricting the ability to use better tissues. The most germane example to the PD situation involves insects that feed on xylem, which generally is thought to contain extremely low concentrations of usable nutrients. Xylem also lacks or contains very low concentrations of allelochemicals (Rosenheim et al., 1996). Xylem-feeding insects usually consume large quantities to extract the limited nutrition (Brodbeck et al., 1993), and they generally cannot exploit other plant tissues, partly because of their specialized mouthparts, but also because they presumably are not adapted to the higher allelochemical concentrations found in more nutrient-rich portions of a plant.
There are fewer examples of insect adaptation to physical deterrents. Plant tissue toughness is a trait that some insects have overcome in ways that sometimes lead to restrictions in host use (e.g., Howard and Giblin Davis, 1997). Surface hairs and trichomes can combine to form to form a physical and chemical barrier, particularly to small insects. The physical structure of some plants can offer unique opportunities for insects to escape predators or parasites, although the degree to which this influences the evolution of diet is debatable (Mira and Bernays, 2002; Scheirs and De Bruyn, 2002; Zangerl et al., 2002). It is generally easier to understand insect adaptation to plants and the reasons for using particular plant species on the basis of chemistry than it is to explain the adaptation on the basis of adjustment to the physical traits of the host plant.
Insects do not always use the same plant species for food, shelter, and reproduction, so noncrop, alternative hosts might exert HPR-based effects on crop pests. There is long-standing controversy about the degree to which plant suitability for feeding adults is linked to suitability for offspring placed there (Wiklund, 1975). Plant traits that influence feeding might or might not be the same traits that influence oviposition or mating behaviors. When more than one plant species is exploited during the life cycle of an insect, each plant and its traits can make separate, significant contributions to the insect’s success.
Insect herbivores can become agricultural pests when conditions permit them to consume or destroy an economically significant fraction of the agricultural product. There are several HPR-based causes of this outcome. One is the tendency to eliminate chemical (or physical) deterrents from food products, because they can be distasteful or toxic to human consumers as well. Another is to bring plant and insect species together for the first time. A lack of coevolutionary history could mean that the plant has no viable defenses against a newly encountered insect. Because insect adaptation to host plants is continuous, uniform populations of crop plants that express the same resistance traits favor the rapid evolution of insects to overcome those traits. For this reason, and because ecosystem complexity itself interferes with finding and exploiting host plants, the simplification inherent to modern agriculture also contributes to the development of insect pest problems.
Sharpshooters and Their Host Plants
Sharpshooters Are Xylem-Feeding Leafhoppers
Sharpshooters comprise a taxonomic subset of a family of insects (Cicadellidae) in the order Homoptera collectively called leafhoppers. All leafhoppers have piercing–sucking mouthparts, and they consume liquid from plants. Some species puncture individual cells or groups of cells, for example, the leaves; others penetrate xylem cells and consume the materials there. Sharpshooters are primarily xylem feeders, although some recent unpublished work suggests that they occasionally feed on cells that surround the xylem.
Sharpshooters and other xylem-feeding leafhoppers have powerful cibarial pumps for extracting xylem contents (which are under negative pressure), and they consume large quantities of fluid, presumably to compensate for the low nutrient concentrations in xylem. There is some uncertainty about where on an individual plant sharpshooters prefer to feed. It is clear that tissues surrounding the xylem, a site of lignification, can become too tough for leafhoppers to penetrate; young tissues would be suggested as preferred feeding sites.
Like many leafhoppers, sharpshooters, temporarily or permanently harbor microbes. They have a diverse assemblage of obligate symbiotic bacteria, the function of which is unknown (Moran et al., 2003). And of course, the plant-pathogenic bacteria they carry can be transmitted from plant to plant. There has been little work on host plant's influences on essential symbiotic microbes in insects or of plant responses to entomosymbionts.
Sharpshooters Have a Broad Host Range
All of the sharpshooter species involved in transmitting Xf appear at first to have large host ranges. They have been associated with many plant
species—native and introduced—including many that grow adjacent to vineyards. This could be important because Xf can be acquired from other plant species and transmitted to grape or citrus and because the abundance and fitness of the insect vector could be influenced by the availability of suitable host plants. The insect’s association with a plant species could mean that it feeds there, reproduces there, or both. There are examples of both types of association with various host plants, but the committee does not have a comprehensive picture of the associations between any of the Xf-transmitting sharpshooters and the host plant communities in their habitats. Indeed, the degree to which any of the sharpshooters actually feed, grow, or reproduce on the locally available plants has been established for relatively few plant species, and not always for plants found in or near California’s vineyards (Brodbeck et al., 1995).
The literature evidences some confusion about which plants are suitable or preferred (Purcell, 1976, 1981; Purcell and Frazier, 1985), and the literature on host plant associations of insect herbivores is notoriously unreliable. An association is often assumed when an insect is found on a plant, although that says nothing about the insect’s performance or preferences. The concept of preference should comprise a quantitative measure of active choice (among equally available alternatives), of choice and performance (growth, reproduction), or at least of abundance on a plant in relation to that plant’s availability. None of those approaches appears to have been taken. Much of what is known about sharpshooters’ use of hosts consists of singular, isolated observations, often of a single insect on a plant. Despite numerous studies in which sharpshooters have been trapped as they move among or within crops and other communities (e.g., Blua et al., 2001), the degree to which the sharpshooters prefer grape or find it suitable for growth or reproduction compared with surrounding vegetation lacks quantitative confirmation.
Factors Influencing Sharpshooter Plant Use
To employ HPR, it is necessary to know what plant traits determine the pest's use of a plant. Few if any of the usual factors thought to influence host plant use have been investigated as barriers to sharpshooter feeding or oviposition on any plant species. There has been little work of this kind on xylem feeders in general, probably because of the difficulty in acquiring and analyzing xylem contents. Studies tend to focus on nutritional chemistry and mostly ignore putative defenses (allelochemicals; Brodbeck et al., 1990,1993,1995,1999; Ponder et al., 2002). Published work links sharpshooter growth to amino acid content and balance in xylem of Glycine max, Lagerstroemia indica and Euonymus japonica, a Vitis hybrid, and Catharanthus roseus (Andersen et al., 1992; Brodbeck et al., 1990,1993,1995,1999). Growth and selection of host plants by another leafhopper, Carneocephala floridana (Ball), was influenced by plant fertilization (Rossi et al., 1996). Grafting on rootstocks that alter xylem amino acid content can influence sharpshooter feeding and growth (Gould et al., 1991) but recent unpublished work failed to
show similar differences in Vitis grafts. Compensatory feeding can confuse interpretation of results in studies like these: Insects can appear to consume more of an unsuitable plant to compensate for its low quality nutrition. If Xf transmission by sharpshooters is linked to time spent feeding, low-quality plants (from the insect’s point of view) could be more likely to become inoculated.
The assumption that sharpshooters can exploit many plant species makes evolutionary sense if the assertion that biochemical challenges (such as defenses) in xylem are few, because host specificity often is assumed to be a product of adaptation to plant exploitation barriers (for example., biochemical defenses, above). Xylem-feeding leafhoppers appear to have adapted to the low nutrient concentrations typical of most xylem, and if there are no other chemical influences, one plant’s xylem thus could be as suitable as another’s. Some xylem-feeding leafhoppers do exhibit relatively low activity of enzymatic detoxification systems and comparatively high susceptibility to chemical pesticides, an observation taken to indicate that they do not typically encounter substances that cause toxic reactions (Rosenheim et al., 1996). Despite the attractiveness of this logic, there are few studies of xylem chemistry or xylem-feeding leafhopper performance to support it. It could be that physical toughness, not chemistry, is the major barrier to leafhopper feeding (Brewer et al., 1986).
But the observations that sharpshooters apparently avoid or perform differently on various plant species (Brodbeck et al., 1992, 1995; Purcell and Frazier, 1985) and that various plant species exhibit genotypic resistance to xylem-feeding and other leafhoppers (Elden and Lambert, 1992; Kornegay et al., 1989; Martinson and Dennehy, 1995; Sanford et al., 1990; Tingey and Laubengayer, 1981) suggest that plant traits do influence sharpshooter attack. Leafhoppers must locate plants or suitable habitats from a distance, determine a plant’s suitability, upon contact, penetrate tissues to get to the xylem, and grow and reproduce adequately as a result of feeding on the xylem. Plant odors, the taste and texture of tissues superficial to xylem, xylem toughness and xylem allelochemistry could produce the observed preferences and hence could be employed in HPR strategies. None of those possibilities has been investigated in any systematic way for Xf-transmitting sharpshooters, although trichomes have been found to constitute barriers to feeding by leafhoppers in several plant species (Elden and Lambert, 1992; Shockley and Backus, 2002; Tingey and Laubengayer, 1981) as has xylem lignification toughness (Brewer et al., 1986).
Plant xylem can include allelochemicals and can express enzyme activity that could influence sharpshooters and provide bases for HPR in grapes. In other plants, alkaloids and other allelochemicals that are toxic to leafhoppers (Sanford et al., 1996) and synthesized in roots are translocated to shoots and leaves in xylem, often in response to aboveground insect attack (Baldwin et al., 1994). Phytoalexins have been found in xylem of several plant species (Kragh et al., 1995; Resende et al., 1996). Xylem of some plants contains phenolics and oxidative enzymes (e.g., Kpemoua et al., 1996; Young et al., 1995). Many authors assume that the presence of those compounds is related to xylem lignification (Kpemoua et al., 1996), but phenolics can be toxic; inhibit enzymes (including digestive enzymes); and could bind to cations, proteins, and amino
acids, especially when activated by peroxidases or other enzymes (Appel, 1993; Morales and Ros-Barcelo, 1997). Phenolic production and activities of oxidizing enzymes can be induced by insect signals (Kruzmane et al., 2002) and could be elevated during pathogen infection in xylem (Kpemoua et al., 1996; Young et al., 1995). Peroxidase activity and phenolic accumulation have been associated with the stylets of feeding insects as they penetrate the plant (Crews et al., 1998) and oxidized allelochemicals form a basis of resistance to leafhoppers in corn (Dowd and Vega, 1996). The discovery of proteases in the saliva of other xylem feeders (Foissac et al., 2002) and unpublished work finding the same in sharpshooters suggests that digestive activity (not just uptake) could occur and could be inhibited by xylem chemistry. Recent work in grapes suggests that rootstock identity—which one might expect to influence xylem chemistry—has little or no effect on aboveground sharpshooter feeding. The xylem chemistry arising from various grafts in Vitis has not been examined.
A promising form of HPR that is consistent with EBPM and IPM is the emission of volatile compounds that attract pest-consuming predators and parasites. There are no studies of grape plant (or other host plant) volatile substances on sharpshooters or their parasitoids, despite the indication that grape leaves produce ample volatile signals when they are wounded by insects (Loughrin et al., 1996, 1997) and that some leafhoppers are responsive to leaf volatiles in other plant species (Saxena and Basit, 1982; Shockley and Backus, 2002; Todd et al., 1990). Recent studies suggest that plants infected by insect-vectored pathogens emit volatile substances that attract the insect vectors (Eigenbrode et al., 2002). There are no studies of the effect of Xf (or other pathogens) on sharpshooter host plant use, or vice versa.
HPR has been developed against several pests in wine grapes, largely by trial and error, by locating apparently resistant material in the field. Vines that are apparently resistant to Xf or sharpshooters and also compatible with producing high-quality wine have not yet been identified by this method. Identifying resistant plants and useful HPR traits would be accelerated and enhanced by understanding how plant and pest interact. A library of phenotypic and genetic resistance traits is needed for effective use of HPR, but few resistance traits have been identified because too little is known about how the insects and plants interact.
Although each of the Xf-transmitting sharpshooter species can be found on and could feed on a large number of plant species, the insects’ preferences for those plants and the suitability of various plants and plant tissues for insect performance have not been determined. It is evident that transmission of Xf to grapevines is linked to insect and disease presence in surrounding vegetation, so it is important to identify with confidence the relative importance of individual plant species in the system.
All plants express heritable, phenotypically plastic “defense” traits thought or known to influence their susceptibility to insects. Sharpshooters and their hosts should be no exception, and the fact that some plants clearly seem unacceptable to them supports this view. Comparisons of plant species’ suitability can provide rough clues about important resistance traits, but similar
comparisons must be made within plant species, because intraspecific variation in resistance provides the raw material for developing resistant plants. And there is informal evidence that such variation exists in Vitis vinifera. But because the basic studies of factors that could influence plant suitability have not been done for sharpshooters, it is difficult if not impossible to guess what sort of plant traits might be useful in developing resistant cultivars or grafts, or in instituting cultivation practices to maximize plant resistance.
Recommendation 3.1. Studies that provide more information about sharpshooter feeding, host-finding behavior, host plant preferences, and the factors that influence reproductive success and natural-enemy-caused mortality are needed. The potential effects of Xf infection on sharpshooter behavior and performance should be included in those studies. Those factors must be examined with statistical rigor so that the results are reliable (Category 1).
Recommendation 3.2. All the modern chemical, molecular, ecological, and statistical tools available to scientists should be used to identify the mechanistic bases of grapevine resistance to xylem-feeding leafhoppers. Studies should be done in the ecosystem context and consider multitrophic interactions among plants, insect pests, and natural enemies (predators and parasites), and they should include insect- and Xf-induced changes in plant quality (Category 2).
Recommendation 3.3. Host plant resistance should be emphasized as a component of ecologically based insect management strategies in the grapevine–sharpshooter–Xf system. Methods for manipulating grapevine resistance should be developed for experimental use to identify key resistance traits and with an eye toward eventual deployment. The methods should allow work with genetically transformed plant material, use of chemical or other elicitors, and cultivation practices (Category 2).
MANAGEMENT OF VEGETATION
Research on plant resistance to insect feeding is complementary to studies of how crop and noncrop components of an agricultural landscape influence the population dynamics and movement of herbivorous insects that attack crops (Gurr et al., 2003) and, if the insects are vectors of plant pathogens, to studies of the epidemiology of those pathogens (A.G. Power, 1990). Vegetation management can be an important element in a comprehensive pest management system and has been effective for the control of insect-transmitted plant pathogens (Wisler and Duffus, 2000). Insects and pathogens respond to vegetation structure and diversity both inside the agroecosystem itself and in the surrounding landscape.
Vegetation Management within the Cropping System
Although the response of individual herbivores to particular cropping patterns varies, crop diversification within an agroecosystem often leads to reductions in herbivorous insect populations (for reviews, see Andow, 1991; Tonhasca and Byrne, 1994; Hooks and Johnson, 2003) and in the pathogens that they transmit to plants (A.G. Power, 1990). The effect of increasing diversity can result either from “bottom up” effects (the effects of the host plants themselves on herbivore population dynamics and behavior) or from “topdown” effects (enhancement of natural enemy populations or behavior). That range of processes which might explain lower herbivore abundance in diverse systems, was first described by R. B. Root (1973) when he posed the resource concentration hypothesis and the enemies hypothesis. The first predicts that herbivores, particularly specialists, in pure, dense host plant stands will be more likely to find their hosts and more likely to survive and reproduce. In contrast, herbivores in less dense or more host-plant-diverse stands should be less likely to find their hosts and more likely to lose them.
Although the details of herbivore–host interactions vary considerably, subsequent experimental tests of this hypothesis generally have supported it with respect to the effects of diversity, especially for specialist herbivores. Host-finding behavior and insect colonization and emigration appear to be important in the response of herbivorous insects to agroecosystems. Densities of specialists could be lower in diverse systems because they have difficulty locating hosts, because of interference with olfactory or visual cues; or because they leave hosts more often because of lower plant quality and then have difficulty relocating the hosts. Those behaviors are significantly affected by the chemical, nutritional, and structural diversity that accompanies crop diversity.
Although the predicted effects of plant diversity on insect herbivores with narrow host ranges are reasonably straightforward, predictions for herbivores with broad host ranges are less clear. The determining factors for insect response are the distribution of host and nonhost plants in the system and the preferences of the herbivore. The expectation of lower abundance in diverse systems attributable to effects on host finding and “host losing” (Kareiva, 1985) is predicated on the existence of nonhost or nonpreferred plants in the system.
The enemies hypothesis predicts that the diverse systems will have higher densities of herbivore natural enemies (predators and parasites) because they provide more resources for those natural enemies, such as alternative prey or hosts, nectar, pollen, and refuge. The hypothesis also has largely been supported in experimental studies, particularly for parasitoids (Andow 1991; Russell, 1989). Compared with monocultures, diverse systems are likely to have higher rates of predation, higher rates of parasitism, and higher ratios of natural enemies to herbivores, all of which can contribute to lower pest densities of specialist and generalist herbivores.
Vegetation Management of Leafhoppers
Leafhopper abundance in crops can be significantly lower in mixtures of host and nonhost species than in monocultures of the preferred host. Those effects have been demonstrated for specialist herbivores—such as the corn leafhopper, Dalbulus maidis (DeLong and Wolcott), whose densities on maize in Central America are significantly reduced in maize and bean polycultures and in weedy maize fields as compared to maize monocultures (A. G. Power, 1987),—and for polyphagous leafhoppers. For example, intercropping with beans reduces colonization of maize by the leafhoppers Cicadulina mbila Naude and C. storeyi China in Africa (Page et al., 1999). Studies in natural grasslands also suggest that leafhoppers are less abundant in diverse plant communities than they are in low-diversity communities (Koricheva et al., 2000).
The most extensive research on controlling leafhoppers with intercropping has been done with the potato leafhopper, Empoasca fabae (Harris), a highly mobile, polyphagous herbivore with an extremely broad host range that can be a serious pest of alfalfa, common bean, soybean and potato. Potato leafhopper populations are significantly reduced when hosts and nonhosts are intercropped. Leafhopper colonization to alfalfa can substantially be reduced by intercropping with oats (Lamp and Zhao, 1993) or with forage grasses (Roda et al., 1997). Compared with bean monocultures, intercropping with tomatoes reduced leafhopper densities on snap beans by 75% (Roltsch and Gage, 1990). Leafhopper populations also were lower on beans in weedy plots than in weed-free plots (Andow, 1992). Populations of potato leafhopper on soybean are reduced when soybean is intercropped with maize (Tonhasca, 1994) or wheat (Hammonds and Jeffers, 1990; Miklasiewicz and Hammond, 2001). Several species of Empoasca are found in lower abundance on squash in polycultures than they are in monocultures (Letourneau, 1990). The use of cover crops or ground covers in orchards and vineyards can help reduce leafhopper populations (Costello and Daane, 2003; Daane and Costello, 1998; Hanna et al., 2003; Roltsch et al., 1998). Surveys and field experiments in raisin grape vineyards in the San Joaquin Valley have shown consistently lower leafhopper populations in vineyards planted with ground covers (Costello and Daane, 2003; Daane and Costello, 1998), although the reasons for the reduction in grape leafhoppers could have more to do with reduction in vine vigor because of competition with cover crops than with such factors as enhanced predation or parasitism. This might point to the need for integration of pest management into viticulture practices.
In many systems, crop colonization by vectors can be reduced by continuous ground cover. In some cases, such decreased leafhopper populations could result from the expansion of predators and parasitoids into the system, although the evidence for it is largely correlative. For example, Roltsch and colleagues (1998) reported greater abundance of predatory spiders and lower abundance of the leafhopper pest Erythroneura variabilis (Beamer) in vineyards with ground cover. Subsequent open-vine exclusion experiments indicated that leafhopper populations were 35% higher when spiders were excluded (Hanna et
al., 2003), although an accompanying spider addition experiment did not show a significant effect of ground cover on leafhopper populations, possibly because of low overall leafhopper numbers (Hanna et al., 2003).
Although no ground cover studies have been done with GWSS, the effects of ground cover on other leafhopper pests of grapes suggest that it would be worth investigating ground covers as one part of an EBPM program to reduce sharpshooter colonization of grape hosts. It will be important to select ground cover species that are not hosts of Xf.
Recommendation 3.4. Detailed, quantitative studies should examine leafhopper performance (survivorship, fecundity, development time) on and preference for a broad range of potential ground cover crops (Category 2).
Recommendation 3.5. The feasibility of using carefully selected cover crops in vineyards to reduce sharpshooter colonization to grape should be investigated (Category 2).
Vegetation Management of Leafhopper-Transmitted Pathogens
Mixtures of host and nonhost species that lead to reductions in leafhopper abundance also can lead to lower prevalence of leafhopper-transmitted pathogens. For example, lower densities of the corn leafhopper in maize and bean polycultures or in weedy maize fields leads to reduced prevalence of the corn stunt spiroplasm in mixtures than is found in maize monocultures (A. G. Power, 1987). Similarly, maize streak virus transmitted by Cicadulina leafhoppers can sometimes be reduced in intercrops of maize with beans or finger millet, although results have not always been consistent (Page et al., 1999). The black-faced leafhopper, Graminella nigrifrons (Forbes), which also transmits the corn stunt spiroplasm to maize, exhibited higher populations in weedy maize fields than in weed-free fields, but the prevalence of the pathogen was lower (Pitre and Boyd, 1970). Feeding preference tests indicated that the leafhopper preferred to feed on some of the weed species, so pathogen transmission to maize was low despite greater vector abundance in weedy systems (Boyd and Pitre, 1969). This example illustrates the complexity of transmission patterns in vector-transmitted pathogens and emphasizes the need to investigate vector behavior in addition to population dynamics in order to elucidate epidemiology.
The transmission of Xf in vineyards could to be influenced by the presence of ground covers, but inadequate information on host relations and plant-to-plant movement of GWSS prevents clear recommendations. To select appropriate candidate species for ground covers, leafhopper performance on, and preference for, a range of crops should be investigated. Candidate species would be examined for water-use requirements (an important consideration in California). Those potential cover crops also must be screened for the capacity to develop epidemiologically significant populations of Xf.
Recommendation 3.6. Potential ground cover crops should be screened for the capacity to develop epidemiologically significant populations of Xf (Category 2).
Vegetation Management outside the Cropping System
The natural vegetation in an agricultural landscape also can have important effects on the population dynamics and movement of herbivorous insects that attack crops. Noncrop plant species can be alternative hosts of insects and pathogens, thereby serving as important reservoirs. Field borders of noncrop plants can increase the populations of natural enemies that visit adjacent crops (Dyer and Landis, 1997; Hickman and Wratten, 1996). Natural areas adjacent to agricultural systems can provide habitat for pollinators and natural enemies of pests (Dennis and Fry, 1992; Landis et al., 2000). For example, Thies and Tscharntke, (1999) demonstrated that rates of parasitism of pest insects and yield losses in oilseed rape were strongly affected by the complexity of European landscapes. In general, parasitism and plant damage were highly correlated with the amount of uncultivated areas, including field margins, fallow fields, grasslands, and forests. Parasitism of the rape pollen beetle by parasitoid wasps in rape fields went from zero in landscapes with little or no uncultivated area to an average of 40% in landscapes with 50% or more uncultivated area. This example demonstrates the potential for natural pest control through modifications in landscape structure.
Vegetation Management of Leafhoppers
There are a few examples of the use of vegetation management at the landscape scale for controlling leafhoppers. The best known is the eradication of Russian thistle to control the beet leafhopper, Circulifer tenellus (Baker), which transmits beet curly top virus to sugarbeets and other vegetables (Wisler and Duffus, 2000). In northern California, the removal of host plants of the blue-green sharpshooter in riparian vegetation bordering vineyards can reduce leafhopper populations (A. H. Purcell, unpublished data).
Vegetation management in the landscape surrounding vineyards could have significant effects on GWSS populations. Studies in Southern California indicate that GWSS is more likely to move into vineyards from citrus groves than from riparian vegetation or nonriparian coastal sage scrub vegetation, particularly in the winter months (Blua and Morgan, 2003). The management of GWSS in citrus groves near vineyards raises difficult issues (see Box 1-1 in Chapter 1), but cooperation between the two agricultural sectors is essential and quantitative and qualitative studies can draw firm conclusions about the role of citrus in the PD problem.
Vegetation Management of Leafhopper-Transmitted Pathogens
The successful control of beet curly top virus was accomplished through the eradication of Russian thistle, a highly suitable host of the virus and its vector, the beet leafhopper, (Wisler and Duffus, 2000). Overgrazing in many western states had allowed native perennial grasses and shrubs to be replaced by the leafhopper’s preferred host plants, including Russian thistle, mustard, and plantain. In Idaho, replacement of the preferred hosts with a nonhost species, crested wheatgrass (Agropyron cristatum), substantially reduced the abundance of beet leafhoppers and virtually eliminated beet curly top as a significant threat to sugarbeet and vegetable production in the 1960s (Wisler and Duffus, 2000). Although current control of the beet leafhopper in California relies heavily on pesticides, good land management practices by livestock farmers to maintain native grass and shrub cover can dramatically reduce leafhopper populations, leading to significant reductions in the prevalence of beet curly top virus (Wisler and Duffus, 2000).
Epidemiologic models suggest that other leafhopper-transmitted pathogens also could similarly be limited by effective vegetation management at the landscape scale. Zhou and colleagues (2002) show that the arrangement of susceptible and nonsusceptible crops in the landscape can significantly influence the epidemiology of the aster yellows phytoplasm transmitted by the highly polyphagous aster leafhopper, Macrosteles quadrilineatus (Forbes). Using a similar modeling approach, Holt and Chancellor (1997) showed that the spatiotemporal pattern of host availability in the landscape can influence the epidemiology of the leafhopper-transmitted rice tungro virus.
The success in reducing abundance of highly mobile, polyphagous leafhoppers through the management of host and nonhost plants suggests that the densities of sharpshooters can also be affected by vegetation management. However, it is essential to understand which hosts serve as feeding hosts and which are reproductive hosts, and it is important to identify the preferences of sharpshooters for those potential hosts. Highly preferred alternative hosts can be sometimes be used as trap crops to attract herbivores away from less preferred crops. This is especially effective for pathogen control where the preferred leafhopper host plant is not a host for the pathogen (Al-Musa, 1982; Farrell, 1976; Toba et al., 1977).
Many riparian plants are hosts of grape strains of Xf in Northern California (Purcell and Saunders, 1999), and some of those species also are feeding and reproductive hosts of the blue-green sharpshooter (Hill and Purcell, 1995; Wistrom and Purcell, 2002). There is also accumulating database on noncrop hosts of Xf in Southern California (Cooksey and Costa, 2003). However, largely because of logistical difficulties, few studies have demonstrated that the pathogen can be acquired from those alternative hosts and transmitted to grape. Cooksey and Costa (2003) indicate that the pathogen can be transmitted from grape to other hosts, such as Spanish broom and wild mustard, but the transmission from those hosts to grape has not been tested. As Wistrom and Purcell (2002) point out, for a plant to be an epidemiologically important source
of Xf, it must be an attractive food plant for sharpshooters and support populations of Xf above 10,000 cfu/g plant tissue, the threshold for acquisition by sharpshooters, after inoculation by an infectious sharpshooter. Many plant species can support some bacterial growth, but relatively few have the characteristics that make them epidemiologically important in the spread of PD. Before undertaking significant weed management programs to control PD, it is essential to collect basic information about the vector’s feeding preferences, bacterial growth potential in alternative hosts, and effective transmission rates from such hosts to grape (Freitag, 1951).
Recommendation 3.7. Detailed, quantitative studies should examine leafhopper preference for potential host plants in the context of natural assemblages of hosts in the field. Studies of leafhopper performance on a broad range of potential host plants are essential to elucidate host ranges (Category 2).
Recommendation 3.8. The plant-to-plant movement of GWSS at multiple scales should be examined throughout the year to identify long-range seasonal and trivial movements that lead to disease spread (Category 2).
Recommendation 3.9. Sharpshooter host plants should be screened for their capacity to develop epidemiologically significant populations of Xf and examined for effective transmission rates from hosts to grape (Category 2).
Recommendation 3.10. After the epidemiologically important noncrop host plants of the vectors are identified, the ecological and socioeconomic barriers to removal of those plants from areas that influence disease prevalence in grapes should be explored (Category 2).
Among the most important tactics and approaches in EBPM are those of traditional biological control. Often, the results of using biological control within an EBPM approach are not as dramatic or as rapid as those produced by pesticide use. There are three primary approaches to the use of predators, parasites, and diseases to manage pests. Classical biological control (importation) involves the collection of natural enemies of the target pest in the country or area from which an introduced pest originated. Introduced natural enemies could establish viable populations (as is typical of classical biological-control introductions) or could require augmentation if pest reduction is inadequate. Augmentation aims to increase the population of a natural enemy (either introduced or native) known to attack a target pest. Typically, this is accomplished by mass rearing one or more natural enemies for field release. Augmentation biological control often relies on continual releases to control. In some cases, augmentation can be more expensive than are other types of
biological control, usually because of rearing, storage, and delivery costs (Obrycki et al., 1997; Tauber et al., 2000), and it can be less cost-effective than are pesticides. However, there have been few if any economic feasibility analyses of augmentative biological-control efforts (Parrella et al. 1992; van Lenteren et al., 1997).
The third approach is the conservation of natural enemies; it involves identifying and then eliminating any factors that limit the survival and effectiveness of natural enemies. For example, native natural enemies that occur in affected areas might not provide necessary control because of the lack of food or shelter for natural enemies. Thus, conservation could be used to enhance biodiversity, and provide food and shelter, thus promoting natural-enemy effectiveness. The economic feasibility of conservation or augmentation biological control is likely to depend on the development of economical, efficient, and effective augmentation strategies or conservation tactics. Leafhopper abundance is regulated by a variety of natural enemies (J. M. Fry 1989). Leafhoppers have a variety of generalist vertebrate predators (birds, lizards) and invertebrate predators (spiders, mirid bugs, wasps, robber flies) (J. M. Fry 1989; Hanna et al., 2003). Leafhoppers also are attacked by various parasitic insects as nymphs or adults (dryinid wasps, epipyropid moths, pipunculid flies, strepsipterans), and as eggs (mymarid, trichogrammatid, aphilinid, and eulophid wasps, although there are only a handful of records of parasitism by species in the latter two families) (Freytag 1985; Heinrichs 1979; Turner and Pollard 1959). Because they feed on plant sap, leafhoppers are usually not susceptible to infection by viral, bacterial, or protozoan entomopathogens. Because entomopathogenic fungi need not be ingested to infect insects, they are the most important pathogens of leafhoppers (Soper, 1985).
The list of important leafhopper pests includes, but is not limited to the grape leafhopper (Erythroneura comes [Say]), the three-banded leafhopper (E. tricincta [Fitch]), the potato leafhopper, beet leafhopper, and the white apple leafhopper (Typhlocyba pomaria McAtee). However, only a few of their natural enemies have been used as biological controls and only with limited or mixed results. Other leafhoppers—Macrosteles phytoplasma, Empoasca vitis Goethe, Scaphoideus titanus Ball, and Scaphytopius magdalensis Provancher—are disease vectors and thus add other dimensions and complexities to biological control. The most commonly used biological-control agents for leafhoppers have been egg parasitoids in the genera Anagrus, Gonatocerus, and Polynema (Flaherty et al. 1985, Freytag 1985, J. M. Fry 1989). However, using Clausen’s (1978) review, Stiling (1994) concluded that establishment of parasitoid species was successful in only 38% of the worldwide biological-control efforts using parasitoids against Auchenorrhyncha. In only 7.7% of the biological-control projects was some degree of control achieved. Reviews of biological-control efforts since 1978 indicate that importations of natural enemies have been done for beet leafhopper, potato leafhopper, and Edwardsiana crataegi (Douglas), (Coulson 1988, 1992, 1994) and success of parasitoid establishment was
inconsistent (Cameron et al., 1989; Clausen 1978, Waterhouse and Sands, 2001). No predators or pathogens have been introduced against leafhopper pests.
CURRENT TACTICS AND STRATEGIES IN BIOLOGICAL CONTROL
Most attempts to manage GWSS using biological control have involved egg parasitoids, most notably those in the Mymaridae and Trichogrammatidae (Triapitsyn, 2003; Triapitsyn and Phillips 2000; Triapitsyn et al., 2002a, b). Exploration for parasitoids has started in Florida, Texas, and Louisiana in the United States and in Argentina, Chile, and Mexico (Triapitsyn and Phillips, 2000; Triapitsyn et al. 2002b; Triapitsyn 2003). Those selected for evaluation as biological-control agents include species reared from GWSS and from other U.S. leafhoppers (Oncometopia nigricans [Walker] and Homalodisca insolita [Walker]) and from exotic species (a South American sharpshooter, Tapajosa rubrimarginata) (Triapitsyn, 2003; Triapitsyn et al., 2002a). More than a dozen egg parasitoid species have been collected.
Most current efforts focus on native and introduced species of egg parasitoids in the genus Gonatocerus, parasitic wasps. Species that have been or are being targeted for importation into California include the mymarids Gonatocerus fasciatus Girault, G. triguttatus Girault, G. morrilli (Howard), G. novifasciatus Girault, G. incomptus Huber, G. atriclavus Girault, and Acmopolynema sema Schauff and the trichogrammatid Ufens spiritus. Most of them have not been propagated successfully; some have been reared in limited numbers. One native species, the egg parasitoid G. ashmeadi Girault, is abundant in California and attacks several leafhoppers including lacerta (Fowler) (smoke-tree sharpshooter), but it is effective primarily during the summer. G. ashmeadi usually causes relatively low parasitism (30-60%) of the spring GWSS egg population (Triapitsyn and Phillips, 2000). Summertime GWSS egg parasitism by G. ashmeadi is higher, often exceeding 95%. The introduced parasitoid triguttatus is receiving a great deal of attention for that application.
Biological Control and Pest Abundance
A major constraint on EBPM of GWSS is its function as a vector of PD. The use of biological control of any pest species with parasitoids, predators, pathogens, or another EBPM tactic, is more problematic when the targeted pest being is a vector (Mahr and Ridgway, 1993). The difficulty arises from the relationship between the size of vector populations and disease spread. Biological control attempts is unlikely to prevent either the initiation of disease in uninfested areas or the spread of the disease in areas where it is established because generally only a few individuals are required to initiate the spread of disease. Nevertheless, the release of biological-control agents or the conservation of indigenous natural enemies (if they are effective in substantially
decreasing vector population) can slow the spread of the disease. Whether biological-control agents do slow the spread of PD should be confirmed with appropriate research. The ability to reduce disease spread, of course, depends on the relationship between the number of effectively transmitting vectors and the efficiency with which disease spreads from plant to plant, within the context of the economic threshold for the particular crop. Those factors are not known for GWSS, PD, and affected crops, or at best they are only now being investigated. Thus, an initial, important question should be asked: Can natural enemies (whether parasitoids or predators) reduce the GWSS abundance?
Biological-control agents generally are assumed to be most effective when they depend on density; that is, if the effectiveness of released biological-control agents increases as pest density increases. Clearly, that assumes that density dependence is necessary for biological control (but see Hanski et al. 1993; Morrison and Barbosa, 1987). Cronin and Strong (1994) reported that in only 25% of the studies they examined was there evidence of density-dependent response in the Auchenorrhyncha. However, they also reported that egg parasitoids were more responsive to host density than were other types of parasitoids. Clearly, any recommendation for the use of biological control as a means of controlling pest populations or even reducing pest populations requires further research into interactions between GWSS and its natural enemies.
Recommendation 3.11. Basic and applied research should establish protocols for the effective selection of natural enemies, develop strategies to increase the success of inoculative releases of parasitoids, and rigorously evaluate the effectiveness of released natural enemies (Category 2).
Basic and Applied Research on Rearing and Release of Natural Enemies
The release of native and introduced parasitoid species is being considered as a component of EBPM for GWSS. The use of introduced parasitoids such as G. triguttatus will likely require the development of rearing, storage, and release technologies. Mass-rearing (which could differ from basic rearing techniques) will be required to augment the population of natural enemies. There are two basic options available for rearing parasitoids or predators: Natural enemies can be reared using GWSS or they can be reared using an alternative host–prey (another alternate leafhopper host for parasitoids or any alternative prey for generalists predators). For either option the host–prey must be reared either on a synthetic diet or on living plants. Although some success has been achieved with live plants, mass rearing GWSS could be problematic. Preliminary data suggest negative consequences for GWSS survival and development when reared at high densities (Lauziere et al., 2002).
One approach to the rearing of GWSS (or another leafhopper) involves the development of a synthetic leafhopper diet and an effective means of providing the diet to feeding insects. The artificial diet presents more problems for leafhoppers than it would for mandibulate insects because leafhoppers have
piercing–sucking mouthparts. So, feeding could require the development of some specialized means of delivering the diet to insects (through membranes, for example, or by sandwich packing of a fluid diet). Some diets have been developed for other leafhoppers (Hou and Brooks, 1975; Hou and Lin, 1979; Wei and Brooks, 1979) but the development of a synthetic leafhopper diet that is appropriate for successful mass rearing is difficult (Mitsuhashi, 1979). There are preliminary efforts to develop such a diet, and any diet that is developed must be relatively inexpensive, considering the number of insects likely to be needed.
Glassy-winged sharpshooters, like other leafhoppers, can be reared on living plants. However there is some doubt about whether a large enough number of herbivores can be produced efficiently that way, and appropriate economic analyses have not been done. Significant improvements in any of the approaches will be needed to produce sufficient numbers of one or more parasitoid species to implement effective, large-scale field releases over the extensive acreage affected by the PD–GWSS. The efforts required to undertake the needed basic research and to implement a mass-rearing program that will produce vigorous and competitive natural enemies are likely to take many years and substantial funding (Neuenschwander et al., 1989).
Recommendation 3.12. Support for classical biological control (inoculative releases) is preferred over augmentation if inoculative releases result in self-sustaining populations and can be shown to be less costly than augmentation (Category 2).
Economic Feasibility of Biological Control
The determination of whether control exerted by natural enemies is sufficient and economically feasible should be made within the context of economic thresholds. There are, of course, several to consider depending on the crop and the type of pest. Economic thresholds are the levels at which control measures are implemented to prevent a pest population from reaching the density called the economic injury level—the lowest population density that will cause economic damage. Another important threshold is the economic damage threshold—the amount of damage level that justifies the cost of implemented controls (Dent, 1991). Thresholds can be used to determine whether single tactics, such as biological control, or a set of EBPM tactics should be incorporated into an EBPM scheme (Reichelderfer, 1981). The establishment of a threshold for a vector species is a daunting task. There currently is no established economic threshold for GWSS, and if such a threshold were to be determined or estimated, it would likely amount to a relatively few individuals. Small numbers of GWSS can transmit enough Xf to cause economically important disease. Development of thresholds is complex and subject to some debate, and even when thresholds are developed they can differ among crops or they can differ depending on the intended use of a crop (wine grapes vs. raisin grapes).
Given what is known about the biological control of vectors, the use of predators and parasitoids to control vector populations is not likely to provide a broadly applicable, effective resolution to the PD problem (Harpaz, 1982). However, only rigorous threshold data will provide an absolute determination of the likely usefulness of biological control. Biological control of GWSS should be viewed as one of many tactics to force abundance of populations to fall below economic thresholds. This presumes that all tactics, in the aggregate, will yield economically important pest reductions.
Determining the economic competitiveness of alternative EBPM tactics is complex and often can be daunting because it could force reliance on assumptions made because of gaps in available data. Determining cost–benefit ratios in circumstances in which partial control is achieved, is perhaps even more difficult. It could be difficult to assign an economic value to the effect of tactics such as biological control, particularly when its use “merely” reduces pest abundance rather than bringing populations below the economic injury level. Many economic assessments of biological-control programs have focused on the value of the crop yield salvaged. In contrast, assessments could determine the economic value of biological control compared with other tactics or strategies (Andres, 1977; Harris, 1979; Hussey, 1985; Zeddies et al., 2001).
“The economic feasibility of any given biological control strategy is unique, and depends on and will vary with the independent and interactive effects of the ecological and economic characteristics of the crop-pest system considered” (Reichelderfer, 1981). No single estimate is representative or indicative of the overall economic feasibility of a biological-control strategy (Reichelderfer, 1981). In general, economic analyses, even when rigorous, can involve assumptions (Headley, 1985) that influence determinations of cost–benefit ratios. They include the amount of control achieved by biological control relative to the use of insecticides (Reichelderfer, 1979), the rate of spread of a pest (Ervin et al., 1983), or the value assigned to environmental quality. Assessments might disregard farmer economic inputs, market dynamics, and other externalities (Reichelderfer, 1981). The factors considered in economic evaluations might differ depending on who the evaluator is (Hussey, 1985). Thus cost–benefit ratios developed for growers, consumers, producers of biological-control agents, or governments could differ in part because each “end user” might choose to include only certain variables or assign different economic value to each factor (Carlson, 1988). End users might include the costs of acquiring a biological-control agent (the costs of foreign exploration, quarantine, production, quality control, storage, transport, release technology) or the costs of alternative tactics, such as the development of insecticides used to control the pest, or the costs of the implementation of alternative tactics. Similarly, the value of improvements in environmental quality, environmental risks and their consequences, or market demand for the crop also varies in different assessments.
There are no economic thresholds to assist managers in PD-affected crops, although methods for their development are available (D.C. Hall, 1988;
Liu et al., 1999; Moffitt, 1986; 1988; Osteen et al., 1988;). All of this points to the need for economic research.
Recommendation 3.13. Research should assess the economic feasibility of biological-control tactics and strategies (Category 2).
Recommendation 3.14. Biological-control tactics within EBPM schemes should be evaluated within the context of working economic thresholds (Category 2).
Compatibility of Management Tactics and Maximizing Effectiveness
There are several constraints on the use of EBPM. In many agroecosystems, even the implementation of IPM currently consists of little more than the application of a variety of insecticides (Ehler and Bottrell, 2000), few of which are benign to the target species’ natural enemies. To achieve the best probability of success, it is necessary to establish conditions that optimize the survival and effectiveness of natural enemies. These conditions would include the reduction of insecticidal drift, and the use of insecticides that selectively target the pest species and are nontoxic to natural enemies (Hull and Beers, 1985). The more accurate the timing of sprays, the more likely that exposure of natural enemies to insecticides will be minimized. Ensuring that these “favorable conditions” for natural enemies persist will likely increase the effectiveness of biological control.
Recommendation 3.15. Research on the use of biological-control agents (predators and parasitoids) should be a priority in commercial vineyards where there is a minimal use of insecticides in vineyards where selective insecticides that are nontoxic to natural enemies are used, or in where the timing of insecticide use is such that mortality to natural enemies is minimal. Similarly, research should be supported that advances the use of biological-control agents in areas or habitats where insecticide use can be severely restricted or eliminated. Areas for study could include riparian habitats, watershed areas, wetlands, and urban and suburban green areas (Category 2).
The committee’s recommendations assume that the objective of biological control is not to prevent the GWSS from vectoring PD but to reduce the GWSS reservoir persisting in nonagricultural areas on noncrop hosts.
Some of the constraints on the use of biological control are the limitations of the organisms themselves or of the environments in which they must exist. Introduced species, if successful, will establish viable populations but could fail to do so at the densities needed for effective pest management. Thus, it might be necessary to augment (when economically feasible) or conserve populations of introduced species or native natural enemies. Because
native parasitoids are species that occur in affected areas but have not provided the necessary amount of control, enhancing their effectiveness is likely to depend on the development of conservation strategies that improve their habitat and provide missing or limited requisites. The habitats to which they are introduced might not be optimal for the natural enemies’ survival and effectiveness and thus, where possible, those conditions should be altered or manipulated to conserve natural enemies, through provision of sugar or nectar sources for adult natural enemies, introduction of alternate hosts or prey, or maintenance of a toxin-free habitat, for example. Similarly, any manipulation, including alterations in viticulture practices, that enhances overwintering survival of natural enemies can be important, particularly because mortality imposed by parasitoids in the spring generation of the GWSS appears to be lower than in summer. Nevertheless, the rigorous experimental confirmation of the importance of such manipulations in agroecosytems has only recently been undertaken, and if manipulations are found to be practical for that agroecosystem, they too should be evaluated experimentally.
Environmental Impact of Biological Control
Many of the consequences of EBPM are unknown. The use of introduced natural enemies, sterile insects, microbial pesticides, and pheromones has sometimes directly or indirectly affected not only the targeted pest species but nontarget plants or other insects. The release of biological-control agents raises ecological concerns, particularly when natural enemies have relatively broad host ranges, as do many of the parasitoids being considered for the use against the GWSS (Triaptisyn et al., 2003). Because many of the parasitoids will parasitize the eggs of other leafhoppers (in the tribes Proconiini, Athysaniini, and others), the effect on leafhopper biodiversity is a valid area of concern that merits investigation. However, important risks also are associated with the failure to control the pests. Thus, concern about the consequences of releasing natural enemies must be balanced against the effect of outbreak levels of GWSS on leafhopper biodiversity or in biodiversity in general. A pest species such as GWSS, which, in California, is a nonindigenous disease vector with a broad hostplant range, can significantly alter unmanaged native ecosystems. The use of introduced biological-control agents has provided significant benefits and fewer risks than have most other control tactics and approaches, although the environmental consequences of biological control are still the subject of considerable debate (Howarth, 1991; Wajnberg et al., 2001).
Human exposure to agents or products used in EBPM can occur at many stages of the production and release of natural enemies. Employees of the facilities in which natural enemies are reared can be exposed to the organisms or to materials used in their rearing. Farm personnel and residents of nearby communities also risk exposure. Although perhaps unlikely, consumers might unknowingly consume microbial pesticides or fragments of arthropod natural enemies in foods. Despite that, few human health effects—with the exception of
allergic reactions—or risks attributable to EBPM agents or products have been reported in the literature. Health concerns arise more often from the use of microbial pesticides.
The recommended incorporation of some tactics and approaches into an EBPM scheme does not mean that other tactics will not be shown to be effective as more research is done. However, available information on leafhoppers in general and on GWSS and its host plants in particular, suggests that the potential of some approaches to the management of the PD–GWSS is highly questionable. In some cases, the absence of data makes an approach impossible to evaluate.
Insect Growth Regulators
Formulations of chitin inhibitors that affect an insect’s ability to develop from one instart to the next have been evaluated. Buprofezin, diflubenzuron and benzoylphenyl were evaluated by Akey and colleagues (2002), and Toscano and colleagues (2002) have studied buprofezin and pyriproxyfen. The chitin inhibitors are applied using standard air blast field sprayers. Akey and colleagues reported that buprofezin controlled eggs and nymphs at a rate of 90–100%, and produced 70% control of adults. Diflubenzuron was nearly as effective, and benzoylphenyl was less effective than the other two inhibitors. Toscano and colleagues reported that buprofezin was effective against first-instar GWSS but that higher doses were required to control third and fourth instars. Pyriproxyfen was effective against GWSS eggs but not against nymphs (Toscano et al., 2003). The effects of chitin inhibitors on beneficial insects were not reported.
One or two formulations that restrain development of chitin have been reported to be effective in preventing nymphs from advancing to the adult stage. Because the nymphs tend to reside on the underside of leaves to avoid predators some new application technology would be required to deposit the spray on those surfaces. Charging fine spray particles can enhance such deposition. However, there are still questions to answer about the amounts of inhibitors required for field control and about the means of application.
Sugar esters, insecticidal soap and neem oil derivatives constitute a new class of insecticides that is being evaluated by organic growers for GWSS control. In one study, sucrose octoanate, sorbitol octoanate, and insecticidal soap were applied with a hand gun sprayer. Several neem derivatives were applied by air blast sprayer. Those materials are being investigated as a means for GWSS
control that circumvents conventional chemical insecticides. Depending on the rate of application, the sucrose materials produced about 98% adult mortality; the insecticidal soap produced 97% mortality rates (Paterka, 2001). The neem derivatives resulted in 33–48% mortality of eggs, 62–78% in nymphs, and 16–50% in adults (Akey, et al., 2002). All of those biorational insecticides are nontoxic to humans and believed to be environmentally neutral.
Research results for biotrational insecticides indicate acceptable control provided by sugar esters and possibly by insecticidal soaps. This is a recent area of research and additional studies would be useful to more clearly develop the application process for such materials. Important questions must be answered: What amounts of these materials should be applied? Do any of them harm beneficial insects?
Biological-Control Agents other than Parasitoids
Biological-control agents of equal or greater effectiveness, such as entomopathogenic fungi, predators, and parasites, could still be developed. The use of entomopathogens such as Hirsutella, a fungus that is known in the southeastern United States to affect sharpshooters, is being explored. However, abiotic constraints, including the need for high humidity, can hamper the use of fungi in many agroecosystems (Soper, 1985). Other options will require considerable basic and applied level research. For example, many of the potentially useful GWSS predators are generalists, and although some of those species have been identified, not all are known. The species that actively feed on GWSS would need to be identified, but more important, their influence on the survival and numerical increase of GWSS would need to be determined. Although many management innovations have resulted from similar research and thus should be encouraged and funded, generalist natural enemies could dramatically affect other indigenous species so it will be important also to study their ecological consequences.
It is not currently practical to use pheromones to cause GWSS mating disruption as part of an EBPM strategy. Although sex pheromones have been used to manage other pests, there is no empirical evidence of pheromone-mediated sexual communication in leafhoppers. Available evidence suggests that mating in most leafhoppers, including species that vector plant pathogens (Claridge, 1985) is mediated by substrate-borne vibrations (R. E. Hunt et al., 1992; R. H. Hunt and Morton, 2001).
Relatively little is known about the host plant chemical compounds that influence the way pest insects find them, begin to feed on them, or initiate other behaviors. The use of behavior-modifying chemicals as attractant lures or repellant sprays is unlikely to result in the effective control of PD–GWSS. As plant and other influences on GWSS host preferences are uncovered, compounds might be identified that could make it practical to use synthesized or natural compounds.
Sterile Male Technique
The sterile male technique has been used to manage pest insects; such as medflies, tsetse flies, and screwworms. Factory-reared flies that have been sterilized, usually by exposure to gamma radiation, are released into a natural population. The males are expected to mate with feral females, thus inhibiting the females’ ability to reproduce. The approach works only if the females mate once, if the released males are competitive with feral males, and if the number of released males at least matches that of feral males. The use of the sterile male technique has been applied against few pest species, in part because of the technique’s biological and ecological requirements. Substantial financial and infrastructure resources also are required to rear and sterilize the millions of individuals generally required. To achieve control, the total pest population in a region must be targeted. Reinfestation of individuals from adjacent areas often interferes with the technique and thus must be prevented with quarantines, insecticidal sprays, or continued releases of sterile males to create a barrier zone that is wider than the flight range of the insect pest. Research has been insufficient, but at present the sterile male technique does not seem to be a useful approach for GWSS management.
The use of mass trapping generally involves traps that are baited with an attractant chemical derived from a living organism. Traps must be distributed over a large area, in a specific pattern. The traps and chemicals are expensive, however, and the need for formulations that are more attractive than are compounds that occur in the ecosystem and the difficulty with the deployment of traps limit the usefulness of the approach. Because there are not data on the use of this approach to control GWSS, the method is not recommended.
Although studies of EPBM techniques could eventually lead to control of PD, it is probable that several years of study with field verification will be required before they are accepted commercially. Until then, grape growers are faced with an expanding infestation that portends serious economic consequences. Research that addresses the use of chemical or biological pesticides offers the most near-term solution to controlling and possibly eliminating the GWSS population in California vineyards and orchards.
Many pesticides can be effective against GWSS eggs, nymphs, and adults (Akey et al., 2002; Toscano and Castle, 2002). GWSS eggs, nymphs, and adults prefer different plants at different stages (Daane and Johnson, 2002). Citrus and grape are highly preferred for egg masses. Oleander and grape are preferred by the nymphs and adults, and adults also favor Agapanthus (African lily). Two peak periods of GWSS egg density on citrus have been identified, one in March and April and the other in July and August (Coviella and Luck, 2003). Coviella and Luck (2003) reported a single peak of nymph activity in May and June and an adult activity in June and July. T. M. Perring and Gispert (2002) reported a peak in adult activity on grape and citrus in June and July. Knowing the preferred locations of eggs, nymphs, and adults and the time of their peak presence can help inform the choice of a control strategy. Several pesticides have been studied for their effectiveness against GWSS at different stages of development. They are discussed by class in the following sections.
Systemic pesticides act through a plant’s vascular system. Neonicotinoids, for example, are introduced into the plant either through the root system or through the leaf cuticle. The pesticides kill the GWSS nymph or adult as it feeds on the xylem fluids. When applied through drip irrigation, the pesticide is taken up by the root tips and transported through the xylem tissue throughout the plant. It takes about 10 days for the compound to reach all parts of a grapevines and about 6 weeks for a citrus tree to become infused (Toscano and Castle, 2002). When applied as a foliar spray, the pesticide diffuses through cell walls and is ingested by GWSS during feeding. There is some evidence that systemics applied to leaves can be diluted by sprinkler irrigation, so that there is incomplete penetration or diffusion (Grafton-Cardwell, 2002).
Applying systemics through drip irrigation systems is an efficient way to treat the whole plant. There would be some loss in the root zone, but worker exposure and drift are minimized. Mortality of 99% of small and large nymphs and 74–79% of adults can be achieved by this method (Akey et al., 2002). Pesticide retention is significant, and data indicate that imidacloprid remains effective in a plant for 1 year (Toscano and Castle, 2002).
Foliar application of systemics can be done by aircraft, by air blast sprayers or by hydraulic sprayers. Aircraft spraying is quick, but it results in
significant drift. Air blast sprayers require a large amount of power, are very inefficient, and produce drift—as much as 256 m after spraying through a canopy in a calm wind (Derksen et al., 2000; Fox et al., 1993). Foliar application of neonicotinoids (where they are translaminar) is effective for a shorter period—3–4 weeks—than is application by drip irrigation, which remains effective for 8–11 weeks (Grafton-Cardwell, 2002). Research comparing the use of air blast spraying with air-assisted electrostatic charging of a fungicide indicated that the dislodgeable foliar residue from grapevines was 68% less with the air blast sprayer, even though the application rate for the electrostatic sprayer was a tenth that of the air blast sprayer (Welsh et al., 2000). This suggests that research should be done to determine whether some of the problems accompanying foliar application of the neonicotinoids can be reduced with a more efficient application technology.
Systemic application is better than foliar application because it requires a smaller amount of the substance to achieve the same plant protection—although that needs verification—because worker exposure and drift are minimized. Research on efficient spray application technologies could lead to reduction in drift for spray applications with ground equipment.
Several questions about the use of the compounds themselves remain, including the quantification of the efficiency of transport into the plant root system and its persistence in the soil after drip irrigation. There also have been reports that repeated use of neonicotinoids, in this case, Assail causes flare-ups of infestatation by spider mites (Michigan State University, 2002).
More serious questions concern the presence of systemic compounds in harvested grapes and their possible consequences for human health. Where systemic pesticides are applied to the surfaces of leaves, what is the role of absorption and adsorption through the cuticle? Their apparently longer term effectiveness warrants intensification of research into the mechanisms of their uptake and movement through the xylem of the plant. More information is needed about their residual presence in the xylem and any effect that might have on grape quality and safety for consumption. Finally, there is some evidence that sprinkler irrigation can dilute or wash out or off systemic formulations. Does this pose a hazard to vineyard workers? When applied by aircraft or air blast sprayers is there a chance that the material will drift into adjacent areas and affect human or animal health? Animal studies have shown neonictinoids are likely to be moderately toxic to humans. They can be highly toxic to birds and bees, although studies show that birds can learn to avoid seeds treated with neonictioids (EXTOXNET, 2004).
Pyrethroids, organophosphates, and carbamates generally are effective but for a shorter period than systemic compounds––3–4 weeks compared with almost 3 months after application (Grafton-Cardwell, 2002). Nonsystemic insecticides are commonly applied by air blast sprayers. Where they are used to
kill GWSS eggs or nymphs it is desirable to place the compound on the underside of the leaves, where the eggs and nymphs reside to avoid predators. Charging small droplets can significantly increase deposition on lower leaf surfaces (Splinter 1968a, b, c). Manufacturers of air-assisted electrostatic sprayers usually recommend application at one-half the rate of active ingredient. There are no reports of spray application equipment used by any of the researchers funded by the PD program other than the air blast sprayers in the sponsored research program reports. Pyrethroids, organophosphates, and carbamates provide close to nearly 100% kill rates for eggs and nymphs and 70%–80% kill rates for adult GWSS (Akey et al., 2002). The are reported to last 2–3 weeks for adults but less than 2 weeks for eggs. Several nonsystemic chemical insecticides have been shown control nymph and adult GWSS, but their effectiveness generally lasts only a few days. In contrast, the systemics show persistence over a growing season. The most effective use of nonsystemic insecticides therefore would be as a rapid response to an invasion of GWSS from adjacent host plants.
Most of those chemicals have exposure limits for field workers and dissipation times before human consumption. There is evidence that carabaryl (Sevin) residue can be dissipated with repeated sprinkler irrigation (Grafton-Cardwell, 2002). Some compounds in the organophosphates group are highly toxic; and chlorpyrifos and dimethoate have been found to be moderately toxic to humans (EXTOXNET, 2004) so there is concern about drift into adjacent residential areas, pasture, or croplands. Organophosphates are readily absorbed by humans and can affect the nervous system through the inhibition of acetylcholinesterase (IOM, 2003; Michigan DNR, 2004).
In addition to determining the toxic effects of those compounds, research should be done to find ways to reduce the amount applied and minimize drift from the point of application. As with any pesticide, efficient application is essential to minimizing the amount of material applied. Conventional application technology—which used air blast sprayers or aircraft—poses inherent problems with the unintentional drift of toxic substances. Commercial application equipment (Law, 1978) is available that effectively charges fine spray particles. The question of whether the amount of pesticide can be reduced through more efficient application methods—while retaining the desired GWSS control level—could be worth investigating.
Finely divided kaolin, marketed as Surround, has been found to repel GWSS nymphs and adults to the extent that they refuse to remain on the plant or to feed (Paterka, 2002; Paterka and Reinke, 2003). A fine coating of this material deters GWSS from settling on treated plants and therefore from oviposition. The material is applied biweekly using air blast sprayers. Kaolin clay has no adverse effects from contact or ingestion.
However, Grafton-Cardwell and Reagan (2003) found that the kaolin coating also repelled the parasitic wasps Aphystus melinus and Comperiella bifasciata and prevented them from attacking California red scale, an unintended consequence to the use of kaolin in citrus. California red scale is not a problem for grapes. The white coating on the leaves, although harmless, raises esthetic concerns that the clay on the grapes could be visually disconcerting to tourists––a major source of revenue for the vineyards––and could taint wine produced from them.
Applying a coating of finely divided kaolin clay has been found to effectively repel both GWSS nymphs and adults. However, the mechanism for this response is poorly understood and the indirect effects (e.g., the impacts on beneficial insects) are also unknown.
ECONOMIC CONTROL STRATEGY
In the short-term, the management of PD is likely to focus on the control of GWSS by pesticides, despite their hazards. The economic sustainability and effectiveness of pesticide use is closely tied to how, when, and where pesticides are applied in the field. For acceptable economic control of GWSS, and to minimize workers’ exposure to the pesticides a precise amount must be applied at the best time, in the correct locations, and for the lowest cost. The factors involved in understanding how pesticides can be used most cost efficiently represent a good case study for the economic analyses needed for all prospective management tools.
Amout of Pesticide Applied
Pesticides are chemical formulations that are intended to kill insects or plant pathogens. Their toxicity to humans is always a concern; even sulfur, which has had a long history of use used to control mildew is being challenged because of its possible allergenicity. And there is always the potential for unintended consequences, as is classically illustrated by the effect of dichlorodiphenyl-trichloroethane (DDT) on the eggshells of birds. Even the inert clay kaolin has met resistance because it whitens grape leaves and because of the concern that it could alter wine flavors. All pesticides carry costs of application, and they all have costs of acquisition.
Therefore fewer applications are better, but the challenge is to determine the minimum amount that will provide acceptable control. The 2002 volume of the Proceedings of the Pierce’s Disease Research Symposium (CDFA, 2002b) contains reports of several studies that evaluate lethal doses for GWSS in a spectrum of pesticides. The work of Blua and Walker (2002), for example, who investigated sublethal doses of neonicotinoids, provides an excellent regression of dose to mortality. That type of study should lead to a rational basis for setting acceptable rates of application.
Timeliness of Application
The best time to control GWSS is before it reaches the adult stage, when it becomes more mobile and can spread the PD as it feeds. Daane and Johnson (2002) reported significant differences in the locations of eggs, nymphs, and adults on host plants. Citrus, gardenia, grape, and hibiscus were preferred for egg laying; oleander and grape were preferred sites for nymphs. Grape and citrus were the primary host plants for adults.
T.M. Perring and Gispert (2002) reported a peak in adult GWSS trap catches in July. If GWSS primarily overwinters as adults, and there appears to be a small population peak in February and March, it would be efficient to schedule a control application with the emergence of the adults at the location of first egg laying on neighboring host plants, such as gardenia and citrus. Grape will be dormant at this time. After the eggs hatch, another application on oleander, hibiscus, and euonymus would target nymphs. Adults should then be targeted in citrus. A second control sequence before the peak in June and July would be indicated, and grapes should be included at this point. Discussions with the research scientists who provided background information to the committee indicate that feeding by GWSS on grapevines after July can infect plants, but that the effect is minimal because the infected stems are pruned. This is reinforced by studies by Coviella and Luck (2003), who reported peaks in egg densities in citrus in March and in August; a peak in the nymph population in citrus was found in June and a peak in the adult population was found in August. They observed high egg mortality during the August in egg density.
Location of Pesticide Application
Conventional application of pesticides has been found to be only 1%–2% efficient for delivering the active material to the effective site (F.R. Hall, 1985; Pimental and Levitan, 1986). This clearly is an area for improvement.
Egg masses generally occur on the bottom surfaces of the host plant leaves. Nymphs also hide on lower leaf surfaces. Mature GWSS evidently feed on larger stems than do smaller leafhoppers. This information should dictate the target area for control.
Toscano and colleagues (2002) reported on simple and effective means of GWSS control that used systemic application of imidacloprid (Admire) through the root system of grapes and citrus. Nymphs and adults were controlled with a single application per season for areawide management programs. The best control was achieved with the application of the pesticide at a rate of 32 oz/acre. The researchers also tested rates of 20 oz and 16 oz/acre. The pesticide moved through the vascular system in 10 days but required about 6 weeks for citrus. Because imidacloprid can be applied through drip irrigation, it has major advantages as a potential control method.
Systemic pesticides can also be applied by aerial or conventional ground rig spraying equipment, requiring only that the active ingredient be
applied to the leaf surface where it can be absorbed. This allows the formulation to be introduced more quickly into the vascular system where GWSS feeds, but one would expect drift losses because vineyards and orchards have large cross-sectional areas that are not occupied by fruit crops.
Many pesticides are applied by air blast sprayers, at 80–100 gal/acre at velocities up to 60 mph. Drift can be measured as far as 100 yd away after spraying through a tree canopy in relatively calm wind (Derksen et al., 2000 Fox et al., 1994). The air blast system uses a large amount of energy, requires the transport of large amounts of solution, and presents obvious concern about drift.
Welsh and colleagues (2000) compared the results from an application of fungicide to grape by air blast sprayer at a rate of 80–100 gal/acre with results from application by electrostatic-spray-charging nozzles at 8–10 gal/acre. They measured dislodgeable foliar residue because they were interested in determining exposure of workers to fungicide. They reported that, at an application of 4 oz/acre a.i., leaf deposition for the air blast sprayer was 68% of that provided by the electrostatic sprayer, as indicated by dislodgeable foliar residue, even though the electrostatic application was one-tenth the volume.
Several current research projects are addressing the effectiveness of a spectrum of GWSS control strategies. The commonly accepted method for pesticide application—the air blast spray system—is inefficient, uses an excessive amount of energy, and creates problems with drift. Newer methods—such as drip irrigation application of systemic pesticides—offer several advantages. The use of air-assisted electrostatic sprayers should be evaluated for application efficiency and drift. Development of that technology for citrus orchards would be appropriate because citrus hosts all GWSS stages. There have been no engineering studies to evaluate the potential for reduction in pesticide application rates or for reduction in drift. Minimizing application rates produces concomitant reductions in cost. Reducing drift addresses a sensitive issue, especially in residential areas. Research is needed to assess the feasibility and effectiveness of methods other than air blast or aircraft spraying, including recirculating tunnel sprays and the charging of fine sprays—as means for improving deposition efficiency and decreasing pesticide drift.
Any new pesticide offers the potential for undesirable or unanticipated consequences. Therefore the potential risk to agricultural workers, to wildlife, to the environment, and to crop quality must be determined as soon as a new control technology is identified. The committee, therefore, makes the following recommendations:
Recommendation 3.16. Control strategies should be pursued that limit the use of insecticides to narrow-spectrum, sustainable formulations that are minimally incompatible with ecologically based approaches to pest
management. A premium should be set on minimizing the negative consequences of pesticide use for human health and environmental quality.
Recommendation 3.17. Research should assess the economic feasibility of specific chemical control strategies and develop decision and cost models to guide growers in setting up chemical control methods for GWSS (Category 1).