Different seasonal feeding activity of diverse herbivores in two temperate forests

The aim of the present study was to understand the effects of abiotic conditions on seasonal feeding activity of diverse herbivores on the same oak tree species in two different forests. We tracked changes in herbivore feeding activities on an oak tree species (Quercus serrata) in two localities: a low elevation small hillock forest patch (Muan, MN) and a middle elevation mountain forest patch (Mt. Jirisan, JR). A total of five sites were selected in each of two forest localities. Data for leaf expansion, leaf chemical qualities, leaf damage ratio, and numbers of lepidopteran caterpillars were collected during spring (May) and summer (July to August), 2012. Leaf expansion rate was higher at the low hillock forest than the mid‐mountain forest from spring to summer. Nitrogen and carbon content decreased seasonally at both localities. Lepidopteran larval diversity was high in the mid‐mountain forest, and two‐way ANOVA showed that species richness of lepidopteran larvae was significantly affected by the interaction between season and locality. Leaf damage by all herbivores was higher in the low hillock forest than the mid‐mountain forest in spring, but was higher in the mid‐mountain forest in summer. Relative proportion of general herbivores increased from spring to summer in the mid‐mountain forest, but not in the low hillock forest. Canonical Correspondence Analysis (CCA) ordination showed that altitude‐ and season‐related variables were significant species and environment interaction factors. Our data indicate that locality and temperature disproportionally affected the feeding activities of diverse herbivores in two different temperate forests.     

Dietary specialization is conditionally associated with increased ant predation risk in a temperate forest caterpillar community

The enemy‐free space hypothesis (EFSH) contends that generalist predators select for dietary specialization in insect herbivores. At a community level, the EFSH predicts that dietary specialization reduces predation risk, and this pattern has been found in several studies addressing the impact of individual predator taxa or guilds. However, predation at a community level is also subject to combinatorial effects of multiple‐predator types, raising the question of how so‐called multiple‐predator effects relate to dietary specialization in insect herbivores. Here, we test the EFSH with a field experiment quantifying ant predation risk to insect herbivores (caterpillars) with and without the combined predation effects of birds. Assessing a community of 20 caterpillar species, we use model selection in a phylogenetic comparative framework to identify the caterpillar traits that best predict the risk of ant predation. A caterpillar species' abundance, dietary specialization, and behavioral defenses were important predictors of its ant predation risk. Abundant caterpillar species had increased risk of ant predation irrespective of bird predation. Caterpillar species with broad diet breadth and behavioral responsiveness to attack had reduced ant predation risk, but these ant effects only occurred when birds also had access to the caterpillar community. These findings suggest that ant predation of caterpillar species is density‐ or frequency‐dependent, that ants and birds may impose countervailing selection on dietary specialization within the same herbivore community, and that contingent effects of multiple predators may generate behaviorally mediated life‐history trade‐offs associated with herbivore diet breadth.     

An altitudinal comparison of caterpillar (Lepidoptera) assemblages on Ficus trees in Papua New Guinea

Aim This analysis of caterpillar (Lepidoptera) beta‐diversity between tropical lowlands and highlands attempts to separate the effects of between‐site (1) turnover of herbivore species on particular host plants, (2) changes in host use by herbivores, and (3) turnover of plant species on changes in herbivore assemblages. Location Two rain forest areas 130 km and 1700 altitudinal metres apart were studied in Papua New Guinea: one in the lowlands (100 m a.s.l.) on the northern coast of the island and one in the central New Guinean cordillera at 1800 m a.s.l. Methods The analysis is based on caterpillar feeding records obtained by quantitative sampling and rearing of caterpillars from four Ficus species studied in the mountains and 21 Ficus species and 62 plant species from other genera and families studied in the lowlands, including three Ficus species studied in both areas. Results Only 17% of species feeding on Ficus in the highlands also occurred in the lowlands. These species represented 1–46% of individuals in caterpillar assemblages on particular Ficus hosts. Widespread species included both Ficus specialists and generalists feeding on numerous plant families. Some of the Ficus specialists changed their preferred host species with altitude. High species turnover was not explained by changes in the species composition of host plants with altitude as lowland and montane assemblages feeding on the same Ficus species showed high turnover. Despite the rarity of widespread caterpillars, the lowland and montane Ficus assemblages were remarkably similar in their dominance structure, species richness, host specificity, generic composition and familial composition. Main conclusions Ficus‐feeding Lepidoptera assemblages between tropical lowlands and highlands are characterized by substantial species turnover not explained by altitudinal changes in the composition of the vegetation. Further, species‐rich plant genera can support caterpillar assemblages with relatively low beta‐diversity compared with species‐poor genera as caterpillars can switch their host preferences from one congeneric host species to another along an altitudinal gradient. Closely related plant species can thus represent a broad, continuously distributed resource along such gradients.     

Density dependence in insect performance within individual plants: induced resistance to Spodoptera exigua in tomato

Net intraspecific density dependence experienced by insect herbivores at the scale of single plants can be a function both of induced resistance in the plant and other interactions among individual herbivores. Theory suggests that non‐linearity in the form of this density dependence can influence the effects of plants on herbivore population dynamics. This study examined both net density dependence at the scale of single plants, and changes in plant quality with herbivore density for Spodoptera exigua caterpillars on tomato plants. One experiment measured the growth of caterpillars moving freely about the plant at different densities, the distribution of damage by these caterpillars, and the quality of the plant as food for caterpillars (growth of caterpillars on undamaged leaf tissue excised from the plant). A second experiment measured plant quality for plants with different amounts of damage by caterpillars confined to particular leaves in mesh bags. Growth of S. exigua caterpillars was found to be negatively density dependent, and this was in part due to decreases in plant quality both as herbivore density increased and as the amount of damage increased. The response of plant quality to herbivores was found to have non‐linear features; there was both a threshold below which no significant decreases in quality (as measured by herbivore growth) occurred, and the decrease in herbivore performance saturated at the highest damage levels. In addition, it was found that caterpillar damage was significantly more aggregated than expected when multiple caterpillars occupy a single plant. This study confirms that host plants have the potential to be a source of density dependence that affects herbivore performance.     

Vertical stratification of a temperate forest caterpillar community in eastern North America

Vertical niche partitioning might be one of the main driving forces explaining the high diversity of forest ecosystems. However, the forest’s vertical dimension has received limited investigation, especially in temperate forests. Thus, our knowledge about how communities are vertically structured remains limited for temperate forest ecosystems. In this study, we investigated the vertical structuring of an arboreal caterpillar community in a temperate deciduous forest of eastern North America. Within a 0.2-ha forest stand, all deciduous trees ≥ 5 cm diameter at breast height (DBH) were felled and systematically searched for caterpillars. Sampled caterpillars were assigned to a specific stratum (i.e. understory, midstory, or canopy) depending on their vertical position and classified into feeding guild as either exposed feeders or shelter builders (i.e. leaf rollers, leaf tiers, webbers). In total, 3892 caterpillars representing 215 species of butterflies and moths were collected and identified. While stratum had no effect on caterpillar density, feeding guild composition changed significantly with shelter-building caterpillars becoming the dominant guild in the canopy. Species richness and diversity were found to be highest in the understory and midstory and declined strongly in the canopy. Family and species composition changed significantly among the strata; understory and canopy showed the lowest similarity. Food web analyses further revealed an increasing network specialization towards the canopy, caused by an increase in specialization of the caterpillar community. In summary, our study revealed a pronounced stratification of a temperate forest caterpillar community, unveiling a distinctly different assemblage of caterpillars dwelling in the canopy stratum.     

Foraging behavior of Cerulean Warblers during the breeding and non‐breeding seasons: evidence for the breeding currency hypothesis

Birds require additional resources for raising young, and the breeding currency hypothesis predicts that insectivorous species exploit large soft‐bodied prey during the breeding season, but shift to small, likely hard‐bodied, prey during the non‐breeding season. To test this hypothesis, we examined prey use by Cerulean Warblers (Setophaga cerulea), foliage‐gleaning Nearctic‐Neotropical migrants, during the breeding and non‐breeding seasons. We collected data on foraging behavior during the breeding season (including observations of prey items fed to young) in upland mixed‐oak forest in southeastern Ohio in 2009 and 2010 and, during the non‐breeding season, in shade coffee in the Cordillera de Merida, Venezuela, in 2008–2009. Cerulean Warblers captured 7% more large prey (visible prey extending beyond the bill) during the breeding than the non‐breeding season, but foraged at similar rates during both seasons. Large, soft‐bodied prey appeared to be especially important for feeding young. We found that adults delivered large prey on >50% of provisioning visits to nests and 69% of identifiable large prey fed to nestlings were greenish larvae (likely Lepidoptera or caterpillars) that camouflage against leaves where they would tend to be captured by foliage‐gleaning birds. Availability of specific taxa appeared to influence tree species foraging preferences. As reported by other researchers, we found that Cerulean Warblers selected trees in the genus Carya for foraging and our examination of caterpillar counts from the central Appalachian Mountains (Butler and Strazanac 2000 ) showed that caterpillars with greenish coloration, especially Baileya larvae, may be almost twice as abundant on Carya than Quercus. Our results provide evidence for the breeding currency hypothesis, and highlight the importance of caterpillars to a foliage‐gleaning migrant warbler of conservation concern.     

Can caterpillar density or host-plant quality explain host-plant-related parasitism of a generalist forest caterpillar assemblage?

Herbivore–carnivore interactions are influenced by the plants on which herbivores feed. Accordingly, dietary generalist herbivores have been shown to experience differential risk of mortality from carnivores on different host-plant species. Here, we investigate whether caterpillar density and host-plant quality play a role in driving variation in generalist forest caterpillar mortality from insect parasitoids using a large-scale, multi-year observational study. A total of 4,500 polyphagous caterpillars were collected from eight host-tree species in Connecticut deciduous forests over 5 years, and frequencies of mortality from insect parasitoids (flies and wasps) were compared across the eight host-plant species for the entire generalist caterpillar assemblage (76 species). Separate comparisons were made using seven numerically dominant generalist species, allowing us to account for variation in caterpillar species-specific parasitism risk. We find significant variation in parasitism frequencies of generalist caterpillars across the eight host-plant species when accounting for variation in caterpillar density. We find no support for an influence of caterpillar density on parasitism and no clear evidence for an effect of host-plant quality on parasitism. Therefore, the results of this study discount the hypotheses that variation in caterpillar density and host-plant quality are responsible for variation in parasitism frequencies across host-plant species. Instead, our findings point to other plant-related characteristics, such as plant-derived parasitoid attractants, which may have robust, community-wide effects.     

Stress magnitude matters: different intensities of pulsed water stress produce non‐monotonic resistance responses of host plants to insect herbivores

Abstract 1. Water stress may increase or reduce the suitability of plants for herbivores. The recently proposed ‘pulsed stress hypothesis’ suggests consideration of stress phenology (pulsed vs. continuous stress) to explain these conflicting effects of plant water stress on herbivore performance. 2. This hypothesis was tested for the effect of differing stress intensity on performance and preference of insect herbivores belonging to different feeding guilds, namely leaf‐chewing insects (Spodoptera littoralis caterpillars) and phloem‐feeding insects (Aphis pomi aphids), on apple plants (Malus domestica). The plants were non‐stressed or exposed to a low or high intensity of pulsed water stress. 3. Plant responses to the different stress levels were generally monotonic. Growth, stomatal conductance (g_s), leaf water, and old‐leaf nitrogen concentration decreased, whereas young‐leaf nitrogen concentration and leaf mass per area (LMA) increased with increasing stress intensity. The stable isotope composition of foliar carbon (δ¹³C) responded non‐monotonically to the drought treatments. The δ¹³C values were highest in low‐stress plants, intermediate in high‐stress plants, and lowest in non‐stressed plants. 4. The preference and performance responses of the caterpillars were also non‐monotonic. Non‐stressed plants were intermediately, low‐stress plants least, and high‐stress plants most attractive or suitable. Aphid population growth was highest on non‐stressed plants and lowest on low‐stress plants. 5. The results highlight the importance of water stress intensity for the outcome of interactions between herbivores and drought‐affected plants. They show that pulsed water stress may enhance or reduce insect herbivore performance and plant resistance, depending on stress intensity.     

Nucleopolyhedroviruses of forest and western tent caterpillars: cross-infectivity and evidence for activation of latent virus in high-density field populations

Abstract. 1. Cyclic population dynamics of forest caterpillars are often associated with epizootics of nucleopolyhedrovirus, but it is not known how these viruses persist between generations or through the fluctuations in host population density. 2. To explore the question of virus persistence at different phases of the population cycle, the nucleopolyhedroviruses of two species of tent caterpillar that co‐occur in British Columbia, Canada, Malacosoma californicum pluviale (western tent caterpillar) and Malacosoma disstria (forest tent caterpillar), were characterised. The cross‐infectivity of the viruses in these two host species was investigated to determine whether there might be a route for virus persistence via the alternative host species. Any virus produced in the cross‐infections was characterised to confirm true cross‐infection or to ascertain whether cross‐inoculation triggered latent virus persisting within the population. 3. The virus associated with forest tent caterpillars (MadiNPV) did not infect western tent caterpillars from low‐density populations, nor did it trigger a latent virus infection; however, inoculation of forest tent caterpillars from high‐density populations with virus from western tent caterpillars (McplNPV) resulted in viral infection, but without a dose–response relationship. 4. Analysis of DNA profiles of virus resulting from cross‐infection of the forest tent caterpillar with McplNPV, revealed that 88% of these infections were caused by MadiNPV rather than McplNPV; however the virus from all 44 infected individuals was identical and differed in DNA profile from the stock MadiNPV used for cross‐infection. This suggests strongly that forest tent caterpillars from high‐density field populations harbour a latent, persistent, or sublethal form of MadiNPV that was triggered by exposure to nucleopolyhedrovirus from the western tent caterpillar. 5. Virus was not activated in western tent caterpillars collected over 2 years of late population decline and the first year of population increase.     

Comparison of Herbivores and Herbivory in the Canopy and Understory for Two Tropical Tree Species1

The Janzen–Connell model of tropical forest tree diversity predicts that seedlings and young trees growing close to conspecific adults should experience higher levels of damage and mortality from herbivorous insects, with the adult trees acting as either an attractant or source of the herbivores. Previous research in a seasonal forest showed that this pattern of distance‐dependent herbivory occurred in the early wet season during the peak of new leaf production. I hypothesized that distance‐dependent herbivory may occur at this time because the new foliage in the canopy attracts high numbers of herbivores that are limited to feeding on young leaves. As a consequence, seedlings and saplings growing close to these adults are more likely to be discovered and damaged by these herbivores. In the late wet season, when there is little leaf production in the canopy, leaf damage is spread more evenly throughout the forest and distance dependence disappears. I tested three predictions based on this hypothesis: (1) the same species of insect herbivores attack young and adult trees of a given plant species; (2) herbivore densities increase on adult trees during leaf production; and (3) herbivore densities in the understory rise during the course of the wet season. Censuses were conducted on adults and saplings of two tree species, raribea asterolepis and Alseis blackiana. Adults and saplings of both species had largely the same suite of chewing herbivore species. On adults of Q. asterolepis, the density of chewing herbivores increased 6–10 times during leaf production, but there was no increase in herbivore density on adults of A. blackiana. Herbivore densities increased 4.5 times on A. blackiana saplings and 8.9 times on Q. asterolepis saplings during the wet season, but there were no clear trends on the adults of either species. These results suggest that the potential of adult trees as a source of herbivores on saplings depends on the value of new leaves to a tree species' herbivores, which may differ across tree species.     

Bird rookeries have different effects on different feeding guilds of herbivores and alter the feeding behavior of a common caterpillar

Aggregations of nesting birds are common in many landscapes and have a broad impact on their surrounding habitat through the nutrient input and disturbance of their guano depositions. Ecological theory makes specific predictions about how disturbances or nutrient pulses will affect the interactions between different trophic levels of organisms. This study dissects the effects of a multi-species bird rookery on plant-herbivore interactions on coast live oak trees (Quercus agrifolia). I found that different feeding guilds of herbivores were affected in opposing directions by the rookery. I observed less chewing damage within the rookery than outside, but more damage from piercing/sucking and galling herbivores within the rookery than outside. To understand why chewing guild herbivores are negatively affected by the environmental impacts of the rookery, I explored the behavioral response of a specialist caterpillar (Phryganidia californica) to various rookery conditions. Bird guano had a direct negative impact on P. californica foraging. P. californica movement was impaired by guano addition to twigs, and caterpillars preferred to eat leaves without guano on them. The rookery also had an indirect negative effect on P. californica foraging, as P. californica preferred clean leaves from forest areas outside of the rookery over clean leaves from within the rookery. This study suggests that while it may be possible to make accurate predictions about the effect of large events (such as rookery formation) on different trophic levels, understanding the response of individual species within that trophic level requires an understanding of aspects of their natural history—such as feeding mode and behavior.     

Preference and performance of Lepidoptera varies with tree age in juniper woodlands

1. As trees age, they undergo significant physiological and morphological changes. Nevertheless, tree ontogeny and its impacts on herbivores are often overlooked as determinants of plant–herbivore population dynamics and the strength of plant–herbivore interactions. 2. Juniperus (Cupressaceae) is a dominant, long‐lived conifer that serves as the sole host to a specialised assemblage of caterpillars. Over the past 150 years, several juniper species in western North America have expanded their geographic occupancy at local and regional scales, which has resulted in an increase in the number of immature trees on the landscape. Using assays in the laboratory, the effects of tree ontogeny on caterpillar performance and oviposition preference for two juniper specialist caterpillars, Callophrys gryneus (Lycaenidae) and Glena quinquelinearia (Geometridae), were examined. The study considered whether responses to tree ontogeny were consistent across caterpillar species and juniper host species. 3. Tree age was found to be a reliable predictor of caterpillar performance, with caterpillars developing more quickly and growing larger when fed foliage from young trees. Differences in the phytochemical diversity between foliage from trees of different ages might help to explain observed differences in caterpillar performance. Interestingly, the specialist butterfly, C. gryneus, displayed an oviposition preference for foliage from old‐growth Juniperus osteosperma trees, despite the fact that larvae of this species performed poorly on older trees. 4. It is concluded that young juniper trees are an important resource for the specialised Lepidopteran community and that tree ontogeny is an important component of intraspecific variation, which contributes to the structure of plant–herbivore communities.     

Life‐history consequences and disease resistance of western tent caterpillars in response to localised,herbivore‐induced changes in alder leaf quality

1. Plants can respond to herbivore damage with phenotypically plastic changes in quality that negatively affect herbivores and prevent subsequent attack – induced defences. 2. The present study tested whether trees respond to herbivory with localised induction, and whether life‐history traits and disease resistance of an insect herbivore are altered on induced branches of the trees. 3. The influences of localised, within‐branch, herbivore‐induced changes in red alder trees (Alnus rubra Bong.) on fitness characteristics of western tent caterpillars (Malacosoma californicum pluviale Dyar) were evaluated. In the field, randomly selected branches of trees were infested with tent caterpillar larvae and the adjacent branches were maintained as non‐infested controls. In the laboratory, larvae were fed leaves from either induced or non‐induced branches through to adult emergence. A second cohort of larvae was challenged with a viral pathogen to compare their disease susceptibility on induced versus non‐induced foliage. 4. Herbivore‐induced, localised responses of damaged branches reduced leaf quality for growth and the fecundity of female western tent caterpillars, but not that of males. Larvae fed induced leaves had a higher survival overall and a reduced mortality due to unidentified non‐viral pathogens than did their counterparts on non‐induced leaves. However, there was no influence of leaf quality on baculovirus‐induced mortality. 5. These findings suggest that localised induced changes in leaf quality could potentially influence populations of tent caterpillars in contradictory ways by reducing their growth rate and fecundity to a modest degree, while improving their survival and resistance to unidentified non‐viral pathogens to a larger extent.     

Grasshoppers amensalistically suppress caterpillar performance and enhance plant biomass in an alpine meadow

Amensalism may be common between non‐trophically linked animals in natural ecosystems, where variation among species in body sizes and foraging modes may give rise to one‐sided interference. However, species and ecosystem‐level consequences of animal–animal amensalism are largely unknown. In a Tibetan alpine meadow, dominant herbivorous grasshoppers trigger a death feigning anti‐predator response of co‐occurring grassland caterpillars despite posing no consumptive threat. We hypothesized that: 1) grasshoppers reduce the performance of caterpillars while incurring no cost to themselves; and 2) this amensalism reduces top–down control of plant composition and biomass. We tested these hypotheses by factorial manipulation of both herbivores within replicate field enclosures. Grasshoppers significantly suppressed caterpillar feeding, growth rate, survival, reproductive effort and delayed metamorphosis. In contrast, grasshopper performance was unaffected by the caterpillars. Suppression of caterpillar feeding decreased overall herbivore suppression of plant biomass by 58% and shifted the functional composition of the plant community (i.e. increased sedge: forb ratio). These results suggest that consideration of non‐trophic interactions such as amensalism will help predict the consequences of species losses for the structure and functioning of ecosystems.     

Vertical stratification of feeding guilds and body size in beetle assemblages from an Australian tropical rainforest

Abstract The vertical stratification of insect species assemblages inhabiting tropical rainforests is well established but few have examined whether these patterns are reflected in vertical stratification of body size or feeding guilds. We used Malaise and Flight Interception Traps to sample beetle assemblages from five locations, at both canopy and ground zones of a tropical lowland rainforest site near Cape Tribulation, Australia. Beetles from 4 years of sampling were sorted to Family and morphospecies, and allocated to one of five feeding guilds. Within feeding guilds the number of species and individuals, from canopy‐ and ground‐caught traps were compared. The body lengths of species were measure and compared within feeding guilds and families. Herbivores was the dominant guild but was not the majority of all species or individuals. Most beetle species (69%) were less than 5 mm in length and the mean size of canopy‐caught species was greater than that for ground‐caught species. This was probably due to slightly more species of plant feeders (herbivores and xylophages) present in the canopy, which were significantly larger than saprophages, fungivores and predators. Among feeding guilds, there were few overall canopy–ground differences. These results contrast with species composition results presented elsewhere where strong differences between the canopy and the ground were evident. We suggest that our guild groupings may have been too coarse to detect fine‐scale differences and that resource partitioning may have also masked faunal stratification. We propose that fine‐scale differences in resources between the canopy and the ground, together with strong microclimate gradients, are likely to be important in structuring the vertical stratification of insect assemblages at the level of species, but not with respect to functional groups.     

Global patterns of arthropod herbivory on an invasive plant,the physic nut (Jatropha curcas L.)

The physic nut (Jatropha curcas L.) is a multipurpose and oil‐producing shrub of Central and South American origin. Since the 15th century, this shrub has existed across tropical regions. Despite its presumed resistance to herbivores, reports show that arthropod herbivores infest it. However, no comprehensive account of arthropod herbivores, which consume the physic nut, exists. Here, we conducted a literature review that provides a comprehensive account of arthropod herbivores of the physic nut. Based on the co‐evolutionary hypothesis, we expected to find a higher herbivore of species richness and a larger proportion of native herbivores within the native range than elsewhere. As physic nut is a well‐defended plant chemically, we expected to find evidence for highest herbivory levels in plant parts that are the least defended. By the literatures review, we compiled 78 arthropod herbivores representing nine orders and from 31 families that feed on physic nut across the globe. As expected, the highest numbers of herbivores (34 species) were documented within the native range of the J. curcas and the lowest species number (21 species) in Africa. Of the 34 species in Central and South America, 94% were of native origin. Nine species were found feeding on J. curcas on more than one continent. Origins of 49% of species were from the native range of J. curcas. The highest percentage (54%) of species belonged to Hemiptera. With regard to feeding guilds, 59% of the herbivores belonged to sucking and 41% to chewing species. Forty‐one per cent of species were flower or fruit feeders, and 36% foliage feeders. We conclude that J. curcas is, despite its toxicity, vulnerable to herbivory, mainly to foliage, flower and fruit feeders.     

How do herbivorous insects respond to drought stress in trees?

Increased frequency and severity of drought, as a result of climate change, is expected to drive critical changes in plant–insect interactions that may elevate rates of tree mortality. The mechanisms that link water stress in plants to insect performance are not well understood. Here, we build on previous reviews and develop a framework that incorporates the severity and longevity of drought and captures the plant physiological adjustments that follow moderate and severe drought. Using this framework, we investigate in greater depth how insect performance responds to increasing drought severity for: (i) different feeding guilds; (ii) flush feeders and senescence feeders; (iii) specialist and generalist insect herbivores; and (iv) temperate versus tropical forest communities. We outline how intermittent and moderate drought can result in increases of carbon‐based and nitrogen‐based chemical defences, whereas long and severe drought events can result in decreases in plant secondary defence compounds. We predict that different herbivore feeding guilds will show different but predictable responses to drought events, with most feeding guilds being negatively affected by water stress, with the exception of wood borers and bark beetles during severe drought and sap‐sucking insects and leaf miners during moderate and intermittent drought. Time of feeding and host specificity are important considerations. Some insects, regardless of feeding guild, prefer to feed on younger tissues from leaf flush, whereas others are adapted to feed on senescing tissues of severely stressed trees. We argue that moderate water stress could benefit specialist insect herbivores, while generalists might prefer severe drought conditions. Current evidence suggests that insect outbreaks are shorter and more spatially restricted in tropical than in temperate forests. We suggest that future research on the impact of drought on insect communities should include (i) assessing how drought‐induced changes in various plant traits, such as secondary compound concentrations and leaf water potential, affect herbivores; (ii) food web implications for other insects and those that feed on them; and (iii) interactions between the effects on insects of increasing drought and other forms of environmental change including rising temperatures and CO₂ levels. There is a need for larger, temperate and tropical forest‐scale drought experiments to look at herbivorous insect responses and their role in tree death.     

Nuclear polyhedrosis virus as a biological control agent for Malacosoma americanum (Lepidoptera: Lasiocampidae)

In addition to damaging trees, the eastern tent caterpillar is implicated in early fetal loss and late‐term abortion in horses. In a field study, we evaluated the potential biological control of the caterpillar using eastern tent caterpillar nuclear polyhedrosis virus (ETNPV), a naturally occurring virus that is nearly species‐specific. Egg masses were hatched and second instar larvae were fed virus‐inoculated foliage to propagate the virus in vivo. Then, a viral pesticide was formulated at concentrations of 10⁴, 10⁶ and 10⁸ polyhedral inclusion bodies per ml. The pesticide was applied to foliage on which second, third and fourth instar caterpillars were feeding. When the majority of surviving larvae reached the sixth instar, colonies were collected and the surviving caterpillars counted. Mean numbers of surviving caterpillars per treatment were compared via 95% bootstrap confidence intervals. The data indicate second instar caterpillars were highly susceptible to the virus, but only at the highest concentration tested. Third instar caterpillars were also somewhat susceptible to high virus concentrations, while fourth instar caterpillars were fairly resistant. Our data provide the strongest evidence to date that ETNPV can be propagated, harvested and refined for formulation as a biological control agent for eastern tent caterpillar. Its use on this insect may be merited in circumstances where landowners and managers need to protect trees and horses.     

Canopy herbivore community structure: large-scale geographical variation and relation to forest composition

Abstract. 1. Geographical distributions of individual foliage‐feeding forest herbivore species have been found to be aligned closely with the distribution of their host trees, however little is known about the extent to which broad herbivore communities are geographically associated with distinct host communities. 2. Large‐scale geographic variation in canopy herbivore communities in a 80 000 km² area (the state of Maine, U.S.A.) was characterised using historical insect survey data. Variation in insect communities was compared with corresponding variation in forest over‐storey composition, which was quantified using data from a regional forest inventory survey. 3. Principal components analysis was used to characterise associations among herbivore and tree species. Analysis of the herbivore data identified three main insect groups: group A corresponded to a single species (Choristoneura fumiferana), group B corresponded to pine‐feeding species including Semiothisa sp., and group C corresponded to the spruce‐feeding species Gilpinia hercyniae and Pikonema alaskensis. Analysis of the forest inventory data characterised three important forest types: northern hardwoods, eastern white pine, and northern spruce‐fir forest types. 4. Spatial analyses were carried out on the first two components of each of the principal components analyses. Factor 1 of the insect data showed a trend of decreasing values from south to north, while factor 2 of the forest inventory data showed an opposite trend. These inverse trends reflected the distribution of the main contributing species to the principal components analysis, C. fumiferana and Pinus strobus respectively. These distributions were highlighted further by the significant negative cross‐correlations that were found between the two factors up to distances of 120 km. 5. Analyses indicated a parallel between the geographic variation in the insect guilds associated with conifers and the geographic pattern of their conifer hosts. Hardwood‐feeding caterpillars, on the other hand, did not form well‐defined guilds and showed varying geographical distributions. The survey data showed varying quality in defining large‐scale associations in the structures of herbivore and host communities. 6. Implications for biodiversity management are discussed.     

Plant‐mediated species networks: the modulating role of herbivore density

1. When herbivores of distinct feeding guilds, such as phloem feeders and leaf chewers, interact, the outcome of these interactions often shows facilitation. However, whether this facilitation turns into competition at stronger herbivory pressure remains unknown. 2. Using an integrative approach that links ecological processes (behavioural choices of insects) with physiological plant mechanisms (nutrient and phytohormone levels) for the wild crucifer Brassica nigra (L.) Koch., this study evaluates preferences of leaf chewers for plants previously infested with several densities of the specialist aphid Brevicoryne brassicae L. (Hemiptera, Aphididae). As leaf chewers, four species of caterpillars (Lepidoptera) were selected that differ in their degree of specialisation in crucifers. 3. These results show that, whereas at low and medium aphid densities caterpillars displayed a preference for aphid‐infested plants or no preference, at high aphid infestation density, all four species of caterpillar preferred uninfested plants, with a significant difference for Pieris rapae and Mamestra brassicae. 4. In contrast to our expectation, the consistent preference for uninfested plants at a high aphid density could not be associated with a decrease in plant nutrition. However, while jasmonate concentrations [i.e. 12‐oxo‐phytodienoic acid and jasmonic acid (JA)] at medium aphid‐density infestation decreased compared with low levels of infestation, at high infestation level, the jasmonates JA as well as JA conjugated with the amino acid isoleucine were present at higher levels compared with low‐infestation treatments. 5. This work provides evidence that positive interactions observed in herbivore communities can be transient, leading to negative interactions mediated by changes in plant defences rather than in plant nutrition.