Workshop on

Spatial Ecology of Plant-Insect Interactions

Tvärminne Zoological Station, Finland
5 - 9 November 2004
 

Abstracts


Oral Abstracts

SPATIAL VARIATION IN THE RELATIVE STRENGTH OF TOP-DOWN AND BOTTOM-UP CONTROL: CAUSES AND CONSEQUENCES FOR PHYTOPHAGOUS INSECT POPULATIONS
Robert F. Denno & Danny Lewis
Department of Entomology, University of Maryland, 4112 Plant Sciences Building, College Park, Maryland 20742, USA.

Most ecologists acknowledge that both natural enemies and host plant resources act in concert to control populations of insect herbivores. What is poorly known is how the strength of top-down and bottom-up control vary spatially and what factors dictate the relative strength of these forces. Here we review factors such as host plant quality, vegetation texture, physical disturbance, and subsidies (allochthonous resources and predators) from neighboring habitats, all of which are known to alter the relative strength of top-down and bottom-up control. Using a salt marsh system as a model, we assemble new data with published information on effect sizes to demonstrate how spatial variation in the above factors integrate to produce landscape-level variation in the strength of top-down and bottom up control. We develop a simple graphical model that integrates spatial variation in top-down ad bottom-up variables to make explicit habitat-related predictions concerning herbivore abundance. The model predicts and field surveys verify that increased disturbance, as it directly and indirectly influences herbivores populations, shifts the balance from top-down to bottom-up control. Based on habitat/vegetation characteristics as well as on variation in the life history traits and behavior of predators and prey we attempt to extend our findings and make general predictions as to the probability for when and where bottom-up and top-down forces will predominate.
QUANTITATIVE FOOD WEBS AND INDIRECT INTERACTIONS:
HOW TO ADD A SPATIAL COMPONENT
H Charles J Godfray & Rebecca J Morris
NERC Centre for Population Biology, Imperial College London, Silwood Park Campus, Ascot Berks, SL5 7PY, UK.

Natural communities contain many interacting species and are spatially extensive. Programmes to understand these two complexities have naturally tended to concentrate on one or the other; for example, studies of the metapopulation structure of one or a few interacting species, or studies of food web structure at single locations. We will begin by describing how the construction and analysis of quantitative food webs (from a single locality) can suggest hypotheses about the processes structuring ecological communities. In a quantitative food web, which typically describes a guild of interacting species, all species and interaction links are expressed in the same units. The hypotheses that emerge can be tested by field manipulation experiments, and we will illustrate this with examples from our work on host-parasitoid communities. We will then discuss how this research programme can be extended to include a spatial component; the logistic challenges that this will entail, and the type of ecological questions it may help understand.
EVOLUTION OF HOST PLANT SELECTION AND DISPERSAL IN METAPOPULATIONS
Ilkka Hanski
Department of Biological and Environmental Sciences, University of Helsinki, PO Box 65, FI-00014 Helsinki, Finland.

In real landscapes, evolution will not be driven solely by local (within-population) processes but also by processes acting at the metapopulation level. The latter processes can be critically influenced by spatial variation in habitat quality and quantity. I will discuss two examples drawn from our long-term study on the Glanville fritillary butterfly (Melitaea cinxia) in the Åland Islands in Finland. First, spatial variation in the host plant composition among the habitat patches influences the rate of immigration of butterflies with dissimilar host plant preferences to these patches. The biased immigration influences the relative use of the host plants in local populations but also the rate of colonization of currently unoccupied habitat patches. The biased colonization rate in turn imposes a metapopulation-level selection pressure that will influence the evolution of host plant preference. Second, the phenotypic composition of local populations in terms of the mobility of female butterflies is influenced by the age and spatial connectivity of populations. Landscape structure will thereby influence the evolution of dispersal rate, and spatial variation in landscape structure becomes a powerful mechanism maintaining variation in dispersal rate and correlated life history traits.
MODELING THE SPATIAL AND TEMPORAL DYNAMICS OF INSECT OUTBREAKS IN A COMPLEX MULTITROPHIC SYSTEM
 Susan Harrison, Alan Hastings & Donald Strong
Department of Environmental Science and Policy, Section of Evolution and Ecology, University of California at Davis, Davis, CA 95616, USA.

To fully understand the complex spatial and temporal behavior of outbreaks, models must go beyond simple plant-herbivore dynamics and consider dispersal, trophic complexity, environmental heterogeneity, and their interactions.  We will examine these factors in models of the bush lupine - tussock moth - ghost moth system. Bush lupines (Lupinus arboreus) in coastal dunes in California undergo periodic mass dieoffs caused by the endophagous ghost moth (Hepialis californicus), which in turn is controlled by a nematode (Heterorhabditis marelatus). The ghost moth is an excellent disperser, but the nematode disperses poorly and is most abundant in areas with high soil moisture, leading to spatial variation in lupine population dynamics. Bush lupines also host outbreaks of the tussock moth (Orgyia vetusta), a flightless defoliator. Tussock moth outbreaks are intense but spatially localized, and experiments and models show this is the result of interactions between the poorly mobile moths and their highly mobile parasitoids (mostly several species of tachinid fly). In the present study, we extend previous models of the tussock moth and parasitoid interaction to include a slowly fluctuating carrying capacity caused by the nematode, ghost moth, and lupine interaction.  By superimposing these two strongly interactive subsystems, we will  generate novel predictions about the qualitative behavior of the whole system.  For example, does the tussock moth population fluctuate "in situ" or does it shrink back to spatial refugia when at low densities? Is the spatial location of tussock moth outbreaks controlled by the nematode and ghost moth interaction?
THE ROLE OF PLANT CHEMISTRY IN REGULATING PHYTOPHAGOUS INSECT POPULATIONS
Erkki Haukioja
Section of Ecology, Department of Biology, University of Turku, FI-20014, Turku, Finland.

Plant defenses may contribute to population levels, and to population fluctuations, of herbivorous insects, and thereby to their spatial patterns. However, time series models have repeatedly indicated that parasitism, not plant defense, do explain changes in population density. I’ll concentrate on two problems, and on their consequences, in the above logic. First, plant defense levels are hard to quantify because plant “defense” is the outcome of three interactive systems: nutrients, secondary compounds, and specific plant defense cascades. Accordingly, there exist no data that would quantify plant defenses in different phases of population cycles. Second, plant quality interacts at least in two ways with parasitism: herbivore-induced plant volatiles attract parasitoids, and the immune defense of defoliators themselves depends on the quality of their host individuals, and on previous damage to these. Short-distance spatial variation in plant quality, or plant defense, is well documented within and among plant individuals. Plant defenses also offer more hypothetical but still plausible mechanisms for large-scale spatial phenomena. For instance, the role of herbivore-induced volatiles in triggering defenses in other plant individuals is well known, but the extent of these mechanisms in natural plant populations is open. Similarly, some environmental factors, like UV-B, switch on large numbers of the same genes which are also involved in specific defense cascades, offering a possible mechanistic basis for large-scale spatial synchrony in population phenomena.
GEOGRAPHICAL VARIATION IN DENSITY DEPENDENT DYNAMICS IMPACTS THE SYNCHRONIZING EFFECT OF DISPERSAL AND REGIONAL STOCHASTICITY
Andrew M. Liebhold1, Ottar N. Bjørnstad2 & Derek M. Johnson2
1 USDA Forest Service Northeastern Research Station, Morgantown, WV 26505, USA.
2 Penn State University, University Park, PA 16802, USA.

Spatial synchrony refers to coincident fluctuations in abundance among spatially disjunct populations. In recent years, there has been growing evidence of spatial synchrony in populations of virtually every type of organism, though the magnitude and geographical range of synchrony may vary considerably. Theoretical models have demonstrated that populations may be synchronized either by small amounts of dispersal between populations or by the actions of small but synchronous random effects, such as that caused by weather (this is the so called “Moran effect”). In previous studies of this type, it has been assumed that populations are primarily determined by density dependent effects and that dispersal and/or stochasticity represents a relatively minor effect. However in virtually all of these previous simulations, it has been assumed that the density-dependent processes affecting populations are identical among various populations.  In this paper we use historical spatially-referenced data on gypsy moth, Lymantria dispar, outbreaks to document that density dependent processes can vary substantially among geographically disjunct populations. This variation may be due in part to geographical variation in the habitat (e.g., variation in forest composition).  We used both linear and non-linear models to explore how among population variation in density dependent processes affects synchronization via either synchronous stochastic effects or dispersal. While results generally support the conclusion that geographical variation in density dependence, serves to diminish synchrony, we were surprised to find that simply the type of density dependence (even when it is identical among populations) can play an equally important role in determining the level of synchronization. Most importantly, the degree to which populations oscillate in a periodic fashion strongly affects synchronization, though this effect varies when sychronization occurs via dispersal vs. when it is caused by stochastic effects. We conclude that studies that attempt to explain synchronization among populations should carefully consider the nature of local density dependent processes and how these processes vary geographically.

PREDICTING LOCAL ADAPTATION IN HERBIVOROUS INSECT POPULATIONS
Susan Mopper
Department of Biology, 300 East Saint Mary Blvd., University of Louisiana, Lafayette, Louisiana 70504-2451, USA.

Generalist insects can become local specialists, even when host plants are sympatrically distributed. Host race formation has evolved in diverse insect taxa and is driven by phenotypic differences between spatially overlapping plant species. The local environment interacts with plant genotype to alter phenology, morphology, defensive compounds, and nutritional status. If sufficiently strong, these selective traits overcome the forces of dispersal and gene flow in an insect population, and create genetic structure shaped by the host species. Compared to host race formation, local specialization on conspecific plants is rare. Nonetheless, demes do form within insect populations, and they occur in insects with very different life history traits. Theory predicted that dispersal ability, feeding mode, and mating system determined the deme formation potential of an insect species. But the empirical data don't support theoretical predictions. In fact, no general patterns have emerged that allow us to forecast reliably in what insect taxa demes are likely to evolve. Perhaps plant rather than insect traits hold the fundamental clues to the evolution of fine scale adaptive structure. I will discuss how wide-spread adaptive deme formation is, what conditions foster deme evolution, and the central role that plants play in driving these fine-scale evolutionary processes.
HABITAT FRAGMENTATION AT DIFFERENT TROPHIC LEVELS
Saskya van Nouhuys
Department of Ecology and Evolutionary Biology, Cornell University, Ithaca, New York, 14853, USA.
Metapopulation Research Group, Department of Biological and Environmental Sciences, University of Helsinki, Finland.

Species experience landscape differently depending on their needs and behaviors. At one extreme are dispersive species, moving at such large scales that they do not perceive a landscape as patchy, or species that experience a landscape as continuous because they have a wide resource breadth. At the other extreme are sedentary species that stay within a patch, and species with extremely narrow recourse needs that are confined to a small fraction of the landscape. Within a community species that interact, even intimately, such as a host and a parasitoid, experience landscape differently. As a consequence, important aspects of the persistence of each species and their interaction happen on different spatial scales. We expect that species at high trophic levels are more sensitive to habitat fragmentation than species at lower trophic levels. But this depends on the relative movement behavior and resource breadth of the species. I will review evidence from empirical studies in the literature, that high trophic level insect species that are sedentary and have a narrow host range are most constrained by habitat fragmentation.
HOST SPECIFICITY AND SPATIAL DISTRIBUTION OF TROPICAL INSECT HERBIVORES
Vojtech Novotny
Institute of Entomology, Czech Academy of Sciences and Biological Faculty, University of South Bohemia, Branisovska 31, 370 05 Ceske Budejovice, Czech Republic.

Analysis of host specificity in tropical insect herbivores is changing from data on insect distribution obtained by canopy fogging and other mass collecting methods to data on insect rearing and experimentally verified feeding patterns. Existing data suggest that a minority of species in herbivore communities feed on a single plant species when alternative congeneric hosts are available. Thus, host plant range limits tend to coincide with those of plant genera, rather than supra-generic taxa. These host specificity patterns lead to downward revisions of previous, extremely high estimates of tropical species richness. Further, they have implications for large-scale spatial distribution of insect herbivores and can lead to the revision of some current assumptions on high beta diversity of tropical insects. Although differences in host specificity between temperate and tropical forests are difficult to assess since data on tropical herbivores originate from recent field studies whilst those temperate derive from regional host species lists, assembled over many years, no major increase in host specificity from temperate to tropical communities is evident.
CYCLIC DEFOLIATING INSECTS IN FRAGMENTED LANDSCAPES
Jens Roland1, Brian van Hezewijk1,2, Subhash Lele3
1Dept. of Biological Sciences, University of Alberta, Edmonton, Alberta, T6G 2E9, Canada.
2Agriculture and Agri-Food Canada, 5403 - 1st Avenue South, P.O. Box 3000
Lethbridge, Alberta, T1J 4B1, Canada.
3Dept. of Mathematical and Statistical Sciences, University of Alberta, Edmonton, Alberta, T6G 2G1, Canada.

The duration of some insect outbreaks has been shown to vary with the spatial configuration of the landscape. Outbreaks of forest tent caterpillar last several years longer in fragmented forests compared to in contiguous forests. Fragmentation is shown to reduce parasitism and transmission of and mortality from viral disease. Using statistical modelling, we assess the effect of landscape on the density-independent, direct density-dependent and lagged density-dependent components of population change, and then relate these patterns to results of field studies on the mechanistic effects of landscape on population processes.
OAKS AS MOSAICS FOR A HOST-SPECIFIC INSECTS
Tomas Roslin & Sofia Gripenberg
Metapopulation Research Group, Department of Biological and Environmental Sciences, University of Helsinki, Finland.

Oaks in the genus Quercus have been classic targets for studies on insect-plant interactions. But how is the local interplay between an oak and its herbivores affected by its spatial setting? And how important is spatial context as compared to intraspecific variation in host plant quality? In this talk, I summarise the results from a series of studies aimed at 1) describing spatial variation in host plant quality, and 2) disentangling the relative effects of spatial context and host plant quality on local insect populations. I report patterns in the physical and chemical attributes of oak leaves across several hierarchical scales, and compare these findings to patterns in the distribution, abundance and performance of oak-feeding insects. I conclude that from the perspective of a specialist insect, individual oak trees form mosaics of highly heterogeneous resource quality, whereas average quality differs relatively little among trees. I end by discussing the implications of these findings for ecological sampling designs, for the evolution of local adaptations and for the metapopulation dynamics of oak-associated taxa.
ISOLATION BY HOST, BY DISTANCE AND BY LANDSCAPE FEATURES COMBINE TO INFLUENCE POPULATION GENETIC DIFFERENTIATION IN CHECKERSPOT BUTTERFLIES
Michael C. Singer & Brian Wee
Integrative Biology, Patterson Laboratories, University of Texas, Austin, Texas 78712, USA.

A cluster of recent studies shows that genetic differentiation of insect populations as estimated from markers of unknown function depends not only on spatial isolation but also on differences in host affiliation. For example, in the checkerspot butterfly Euphydryas aurinia there are significant effects of distance independently of host and of host independently of distance. However, there is an interaction between the two: host effects explain a high proportion of the genetic variance when two different hosts are used allopatrically but a low proportion when the same two hosts are used sympatrically. The nature of the interplay between effects of physical separation and local adaptation seems at present to differ dramatically among insect species and even among groups of populations that use particular hosts. For example, using AFLP analysis of nuclear DNA in the checkerspot butterfly Euphydryas editha, we showed that populations in which insects feed on Castilleja had high isolation by distance (IBD) and strong among-population genetic differentiation. In contrast, populations where Collinsia is the host showed no IBD and low genetic differentiation, measured over a similar geographic area of about 1000km x 300km. This difference may support the suggestion from a separate analysis of mtDNA that Castilleja is the original host of this insect species. Puzzlingly, a superficially similar effect, in which insects on Castilleja have high genetic variance, occurs at an entirely different scale of 300m x 300m. Inter-individual variation was significantly higher among larvae gathered from Castilleja compared with those gathered from Pedicularis at the same site, although there was no difference at all in mean insect genotype on the two hosts. Larvae from Castilleja also had significantly higher heterozygosity as individuals than those from Pedicularis. Butterflies at the site didn't have strong preferences for either host species so we suspect that the differences we measured were not present at oviposition but developed as a result of differential larval mortality during the first few days of feeding.

COEVOLUTIONARY MOSAICS
John N. Thompson

Department of Ecology and Evolutionary Biology, Earth and Marine Sciences Building , University of California, Santa Cruz, CA 95064, USA.

As species coevolve, they adapt and coadapt in different ways within different populations. Local coevolution produces a small range of coevolutionary dynamics that have now been well characterized in theoretical models and demonstrated in laboratory and natural populations. These local dynamics form the modules upon which the broader coevolutionary dynamics of species take place. As our understanding of the geographic mosaic of coevolution has grown in recent years, it has become possible to develop more specific hypotheses and predictions on how pairs and groups of species coevolve across complex geographic landscapes.

SPATIAL SCALE OF LANDSCAPE CONTEXT INFLUENCING LOCAL INTERACTIONS AMONG INSECTS, PATHOGENS AND PLANTS
Teja Tscharntke
Agroecology, University of Göttingen, Waldweg 26, D-37073 Göttingen, Germany.

Interacting populations driving ecosystem processes depend on the recruitment of organisms and species from source populations in the surrounding landscape. In simple landscapes, local stands may not receive the set of species necessary to realize the potential ecosystem functioning. As populations of interacting species are influenced at different spatial scales, which may be measured as different radii around the local population, structural changes from small to large landscape radii will affect species differently. Here, I focus on case studies from our group on populations of plants, pathogens, phytophagous insects and their parasitoids as well as plant-pollinator interactions. Patterns of plant diversity are mostly related to local site conditions, but landscape context may also be important. This is shown here by analysing the relative importance of local site and regional landscape characteristics on local species richness of arable weeds. More details will be given for Creeping Thistle Cirsium arvense populations. Changes in local plant densities with landscape context affect the resource basis for trophic interactions. Levels of local rust infection of Creeping Thistles appeared to be determined by larger scales of the surrounding landscape than the thistles’ populations. In addition to an overview on thistle insects and their relation to pathogens, landscape context of cereal aphid-parasitoid interactions is given showing that these host-alternating aphids as well as their enemies profit from complex landscapes. Analyses of local plant-pathogen-insect interactions need a landscape perspective, but we are far from understanding the complex mechanisms driving populations of interacting species in real landscapes.

COMBINING MODELS AND DATA IN ANALYSING HERBIVORE-HOST PLANT INTERACTIONS
Eddy van der Meijden & Sonja Esch

Institute of Biology Leiden, Leiden University, The Netherlands.

During the seventies of the last century we started a project on a (monophagous) insect herbivore-host plant interaction with periodic total defoliation. The aim was to test whether the dynamics of the system would fit a simple Lotka-Volterra oscillation model. Within two years, two out of three selected local host plant populations, and consequently those of the herbivore, became extinct. We realized that the system could only survive in a spatial configuration. Since then one hundred local populations of insect and plant have been monitored. Ignoring the spatial configuration, on the metapopulation level herbivore and host indeed can be described by an oscillation model. Delay of recovery of the herbivore after large-scale metapopulation crashes, a delay which is probably crucial for host plant recovery, cannot be explained by food shortage. This delay runs parallel to relatively high parasitoid incidence. Models indicated that either a better dispersal capacity of the herbivore or a larger fecundity led to new outbreaks during which parasitism is reduced to almost zero. These models guided experiments on parasitoid dispersal and on a molecular comparison of herbivore and parasitoid populations in a spatial context. Finally we will analyse herbivore-host plant interactions in spatially separated systems in different countries with and without parasitoids.
METAPOPULATION DYNAMICS AND GENETIC STRUCTURE OF MONOPHAGOUS INSECT HERBIVORES
Wolfgang W. Weisser
Institute of Ecology, Friedrich-Schiller-University,  Jena, Germany.

In patchy environments, patterns of genetic variation will be strongly influenced by migration between local populations and by local colonization and extinction events. The aphid species Metopeurum fuscoviride and Macrosiphoniella tanacetaria are specialised herbivores on tansy, Tanacetum vulgare, a perennial plant common along rivers, the coast and in wastelands. Due to clonal growth, tansy plants (=genets) consist of one to many shoots (=ramets), each of which can host an aphid colony. The habitat of tansy aphids is therefore highly structured and aphids cluster in at least three spatial scales: at the level of ramets, at the level of genets and at the level of sites. (Sub)population turnover at all these levels is very high, with persistence times in the order of weeks (ramet) to months (sites), suggesting that migration rates should also be high, resulting in little differentiation between populations. However, there is high genetic diversity at all spatial levels and populations are significantly differentiated, even over a small spatial scale. While the ecological data appear to show that most dispersal occurs within sites with little between-site migration, the genetic data suggest that founder effects and genetic drift in small populations appear to dominate the genetics of the metapopulation.


Poster Abstract

SCALE-DEPENDENCE OF PLANT-HERBIVORE INTERACTIONS: BARK BEETLES AND PINE TREES
Matt Ayres1, Tiina Ylioja2, Sharon Martinson1 and Brenda Whited1
1 Dartmouth College, Hanover, NH, USA
2 Metla, Finnish Forest Research Institute, Vantaa, Finland.

When and why do ecological patterns change character with spatial scale? We have begun exploring some surprising changes across scale in the pattern of interactions between bark beetles and their host trees. We compared the susceptibility of Pinus taeda and P. virginiana to Dendroctonus frontalis (Coleoptera: Scolytidae) at two hierarchical scales: within heterospecific stands (ca. 10x10 m) and across a forest landscape (72 000 ha) that was dominated by monospecific stands. In the former, flying beetles preferentially attacked P. virginiana (tree mortality = 65-100% in P. virginiana vs. 0-66% in P. taeda), but in the latter, stands of P. taeda were more susceptible than stands of P. virginiana. This hierarchical transition in host susceptibility was predicted from knowledge of (1) a behavioral preference of beetles for attacking P. virginiana vs. P. taeda, (2) a negative correlation between host preference and reproductive success, and (3) a mismatch in the domain of scale between demographics and host selection by individuals. We have identified two additional incongruencies that arise with changes in spatial scale: 1. Bark beetle populations display strong positive density-dependence (Allee effects). The orthodox explanation has been that large aggregations of beetles are more successful in overwhelming tree defences. This is true for D. frontalis, but there turns out to be no relationship between local population size and average attack rates of individual trees. Thus, reproductive benefits of mass attacks on individual trees cannot readily explain the higher relative growth rate of large infestations. 2. Single species stands of Pinus palustris are strongly resistant to D. frontalis infestations even though individual longleaf trees within mixed species stands do not seem any more likely to escape attack or survive attacks than loblolly trees. Solving these puzzles should yield an improved understanding of population dynamics in D. frontalis, and could aid in developing a general approach for recognizing and understanding scaling allometries in the demographics of other herbivores.
THE IMPORTANCE OF SPATIAL SEARCH BEHAVIOUR
Matthew Collett
Department of Zoology, Michigan State University, USA.

How is a herbivore population affected by changes in the abundance of its host plant? If the herbivore population is controlled by predators or parasitoids which move frequently between host-plants, then the answer depends crucially on the fine-grain spatial distribution of the host plants and the spatial search behaviour of the predators. Changes in the abundance of host-plants will have a non-linear effect on herbivore abundance that can be described in terms of a threshold of host-plant abundance. The size of this threshold is determined by the relative speeds with which the predator moves on host-plants and in the areas between the plants. As long as host-plant abundance remains above the threshold, then changes in host-plant abundance will have little impact on the number of herbivores on each plant. Below the threshold, however, a reduction in host-plant abundance will result in an increase in the number of herbivores on each plant. And the less direct the predator’s travel between plants or patches of plants, the more herbivore abundance will increase. These results suggest that the spatial abilities of individual predators and the spatial habitat structure modulate the relative strengths of top-down and bottom-up forces in controlling a herbivore population.
GENETIC PATTERNS OF DIFFERENTIATION IN THE LARGE PINE WEEVIL IN EUROPE
C. Conord, L. Despres & G. Lemperiere
LECA : Laboratoire d’Ecologie Alpine, Université Jospeh Fourier - Grenoble 1, France.

The large pine weevil, Hylobius abietis, (Coleoptera: Curculionidae) is a major economic pest of reforestation crops in Europe. Little is known on its precise dispersion ability and population genetic structure, yet this knowledge is crucial for establishing control strategies. Damages on seedlings are due to adults; larvae develop on fresh stumps of Scot Pine and Norway Spruce. The population dynamics of H. abietis strongly depends on the patchy distribution of clear cut areas (breeding sites) and of restocking sites (maturation feeding of the adults), as well as on the available host-plant species. We developed and used Amplified Fragment Length Polymorphism markers (AFLP) to assess genetic structure at the continental and at the regional levels by sampling populations throughout Europe (Estonia, Finland, France, Ireland, and Poland). We also looked for evidence of genetic differentiation according to the host-plant using larvae sampled from stumps, either in mixed or in monospecific conifer plantations. Although weak, a statistically significant geographical structure was found at the European level. However no difference was found between larvae developing on Scot Pine and those developing on Norway Spruce. Large scale adult migration is likely to overcome the formation of two host-races, preventing local adaptation. The weak genetic structure observed throughout Europe may favour the development of efficient biological control strategies.
CONNECTIVITY AND PATCH SHAPE CAUSE DIFFERENTIAL SEED PREDATION
Ellen I. Damschen1 & John L. Orrock2
1Department of Zoology, North Carolina State University, Raleigh, North Carolina, USA.
2Department of Ecology, Evolution, and Organismal Biology, Iowa State University, Ames, Iowa, USA.

Increasing connectivity between habitat patches is thought to mitigate negative effects of habitat fragmentation for plants by promoting pollination and dispersal. These positive effects may be offset, however, if seed predation, herbivory, or disease is also increased. In a well-replicated experimental landscape that tests for the effects of connectivity (via landscape corridors) while controlling for patch area and shape, the deposition of bird dispersed seeds has been shown to be higher in connected vs. unconnected patches. Here, we used exclosures in this same study system to test for differential seed predation on two bird-dispersed plants: black cherry (Prunus serotina), and blackberry (Rubus allegheniensis). P. serotina has large seeds (0.068-0.084g) and was almost exclusively consumed by rodent seed predators, whereas R. allegheniensis has relatively small seeds (0.002-0.003g) and was consumed by both rodents and arthropods. Consumption of both species was significantly greater in connected vs. unconnected patches, suggesting that the benefits of increased dispersal could be negated by the negative effects of increased seed predation. Additionally, corridor-mediated changes in seed predation also differed between the two species. Connectivity may differentially affect the survival of seeds, which could have important consequences for the composition of the seed bank and resulting plant community.
HOST PLANT QUALITY AND LOCAL ADAPTATION AS DETERMINANTS OF THE DISTRIBUTION OF A GALL-FORMING HERBIVORE
S. P. Egan1,2 and J. R. Ott1
1Department of Biology, Texas State UniversitySan Marcos, 78666, USA.
2Department of Biological Sciences, Vanderbilt University, Nashville, TN 37235, USA.

Herbivores exhibit a universal pattern of patchy distribution within populations of their host plants. We report on a manipulative transplant experiment that links host plant quality to herbivore fitness at the local scale with the process of local adaptation at the patch scale to provide a synthetic explanation to this pattern for the gall-forming herbivore, Belonocnema treatae, within populations of its host plant, Quercus fusiformis. Oak/gall-former systems support species-rich assemblages that offer model systems to study community ecology. Thus, a mechanistic explanation for the distribution of a gall-forming insect provides powerful insight into understanding the distribution of entire terrestrial communities. Mated B. treatae females from each of five high gall density trees were allowed to oviposit and their progeny subsequently develop on (a) their respective four nearest conspecific neighbors exhibiting low gall density, (b) the four alternative high density trees, and (c) their natal trees. Mixed-effects ANOVA was used to analyze seven independent estimates of insect performance. Results demonstrate, for the first time in a Cynipid -Oak system (>2 000 species), that (a) stable herbivore populations are isolated by low quality host trees and (b) local adaptation has taken place at the scale of individual trees. Thus, host plant quality and local adaptation act, in combination, as a driving force on the distribution of B. treatae and its diverse gall-dependant community (N = 27 species).
SPATIAL VARIATION IN BIOTIC INTERACTIONS AND FLORAL POLYMORPHISM
Johan Ehrlén
Department of Botany, Stockholm University, S - 10691 Stockholm, Sweden.

Floral characters may influence the attractiveness to pollinators, but also the risk of damage by seed predators and herbivores. Spatial variation in both mutualistic and antagonistic biotic agents may thus result in variable selection on plant reproductive traits. We currently investigate how biotic interactions influence selection on scape length in 52 populations of the perennial herb Primula farinosa, which is dimorphic for scape length and produces either a regular or a very short scape. Fruit initiation is lower and more variable in short-scaped than in long-scaped plants, suggesting pollinator-mediated selection for long scapes. However, the long-scaped morph is more damaged by grazers (cattle, sheep and horses) and seed predators (the moth Falseuncaria ruficiliana) than the short-scaped morph. Variation in the magnitude of morph-specific differences in fruit initiation and damage from grazers and seed predators result in significant variation in the relative fecundity of the two scape morphs among populations. In many populations, short-scaped plants have a higher fecundity, suggesting that antagonists reverse the selection exerted by mutualists. Spatial variation in selection on scape length in P. farinosa is thus determined largely by the intensity of three biotic interactions, and the observed selection mosaic is the integrated result of these interactions.
RELATIVE IMPORTANCE OF RESOURCE SIZE AND ISOLATION FOR LANDSCAPE DISTRIBUTION OF TWO MONOPHAGOUS BUTTERFLIES
Jochen Krauss
Institute of Environmental Sciences, University of Zuerich, Winterthurerstrasse 190, 8057 Zuerich, Switzerland.

Patch size and isolation of host plants are major causes of species extinction. We tested the effects of food plant availability, habitat area and quality as well as habitat isolation for the landscape distribution of the two monophagous butterfly species Cupido minimus and Polyommatus coridon. Both butterflies and their larval food plants are specialized on fragmented calcareous grassland. We surveyed all known calcareous grasslands (n = 298) around the city of Göttingen (Germany) to map the occurrence and population density of the host plants and the two butterflies and recorded habitat area, different habitat quality factors and distance to the next conspecific population of each habitat. Both butterfly species were highly affected by larval food plant availability, which was positively correlated with habitat area. Habitat isolation (up to 5 km for C. minimus and 7 km for P. coridon) and habitat quality played a minor role for landscape occupancy and population density of both species. These factors may be often overestimated, as they only appear to contribute to landscape distribution in highly fragmented landscapes, where these factors shift towards extremes. Hence, general recommendations for conservation programs are difficult.
FOOD PREFERENCE AND PERFORMANCE OF A SPECIALIST HERBIVORE: SPATIAL VARIATION IN HOST PLANT QUALITY
Roosa Leimu & Marianna Riipi
Section of Ecology, University of Turku, Finland.

The generalist/specialist hypothesis predicts that, unlike generalist herbivores, specialists may be unaffected or attracted to defence compounds, and can even prefer higher levels of certain chemicals. Abrostola asclepiadis is a specialist folivore of a perennial herb, Vincetoxicum hirundinaria, which contains toxic alkaloids that deter most generalist herbivores. The food choice, leaf consumption and larval growth of A. asclepiadis were studied to determine whether there is variation among host-plant populations in suitability for this specialist herbivore. There was significant variation in food preference and leaf consumption among host-plant populations, but no differences were found in larval growth among populations. A. asclepiadis larvae preferred host-plant populations with higher alkaloid concentrations, as predicted by the generalist/specialist hypothesis, but consumed less leaf material from plants originating from such populations in a no-choice experiment. Our results suggest that the larvae may benefit from higher alkaloid concentrations, possibly through sequestration for their own defence. Although the growth of the specialist herbivore was not affected by host plant quality, larval food preferences may lead to selection against host plants with high contents of defensive compounds. Ultimately, the opposing selection pressures from specialist and generalist herbivores on plant chemistry may lead to spatial differences in the potential co-evolutionary dynamics between the plant and the herbivores, depending on the types of herbivores present in the environment.
MODELLING VECTOR DYNAMICS AND THE SPREAD OF INSECT-TRANSMITTED VIRUSES
D. Morgan & Hazel Parry
Central Science Laboratory, York, YO41 1LZ, UK.

Barley yellow dwarf virus (BYDV) is an important disease of cereals and grasses throughout the world causing significant losses to crops. In the UK introduction of the disease into a field (primary infection) results from the immigration of viruliferous winged (alate) aphids into fields and subsequent spread (secondary spread) occurs as infectious wingless (apterous) offspring disperse through the crop. In the UK pesticides tend to be applied routinely to control aphid vector populations and irrespective of the risk of virus infection such that unnecessary and ill-timed application may be made. A reliable forecasting system is needed based upon thorough understanding of the biological processes involved but until recently little was known about the factors which determine the introduction and spread of BYDV. Thus the aim of the project was to develop a computer model which predicted the spread of BYDV and which would underpin a rational Decision Support System (DSS) for both vectors and virus alike. A stochastic individual-based simulation model has been developed integrating sub-models describing the population dynamics and behaviour of aphid vectors, and epidemiology of the virus. The system utilises an approach so that each individual aphid and plant within the field is monitored by the model, which utilises routines to simulate aphid development, reproduction and mortality, and the dispersion of vectors between plants and subsequent spread of the virus throughout the crop. The model was validated with observed data of overwintering aphid populations and virus incidence from field experiments in 1996/97, 1997/98 and 1998/99. The predicted spatial dynamics and temporal incidence of both aphids and the virus was similar to those observed in cereal fields. Furthermore the model has provided useful insights into understanding the complex interactions between biological processes involved in BYDV epidemiolgy. A sensitivity analyses of the model identified (1) virus latent period, (2) overwintering aphid mortality, (3) the number of infectious alate aphid migrants, and (4) the dispersal rate of apterous aphids within crops, as critical factors, which have a greater impact on the model output than other factors. Small changes in the associated equations produced a halving or doubling of the final amount of BYDV predicted.
THERE'S NO SUCH THING AS A FREE LUNCH: THE COST OF DEFENCE CHEMICAL SEQUESTRATION IN A LEAF BEETLE
Russell Naisbit
Laboratoire d'Ecologie Animale et d'Entomologie, Université de Neuchâtel, Switzerland.

Plant defence chemistry often backfires, when insect herbivores take those compounds for their own protection. In the alpine leaf beetle Oreina elongata (Coleoptera: Chrysomelidae) two types of defence are known: adults and larvae synthesise cardenolides when feeding on Cirsium spinosissimum, but can also sequester pyrrolizidine alkaloids (PAs) from three species of Adenostyles. Sequestration, by taking advantage of chemicals already produced by the plant, might provide the cheaper strategy. However, there may be costs associated with uptake, concentration, and avoidance of toxicity. Here we test for a cost of chemical sequestration in larvae. We used individuals from six isolated populations, two of which never encounter Adenostyles, and hence PAs, in nature. Larvae were reared on Cirsium, with the addition of either a methanolic solution of PAs extracted from Adenostyles alliariae, or methanol alone as a control. Pyrrolizidine alkaloids significantly reduced growth rate but left development time unaffected, resulting in a 6% reduction in weight at the start of the fourth instar. Apparently, sequestered chemical defence is not free. The effect was consistent across populations, so that like oviposition preference and larval performance, the cost of sequestration provides no evidence for strong local adaptation to differences in host plant availability.
IMPACT OF SHOOT HERBIVORY ON A PERENNIAL HERB IN A METAPOPULATION CONTEXT
Viktoria Pettersson, Hans Gardfjell, Lars Ericson & Barbara E. Giles
Department of Ecology and Environmental Science, Umeå University, 901 87 Umeå, Sweden.

Silene dioica is a dioecious perennial herb and food source for two host-specific, shoot-eating herbivores. Larvae of the microlepidoptera Caryocolum viscariella (Gelechiidae) attack reproductive shoots as they form, spinning together the upper leaves and eating the buds and upper parts of the floral stem. Larvae of the root-maggot fly Delia criniventri (Anthomyiidae) tunnel through the basal meristem to eat into the bases of the flowering shoots. In the Skeppsvik Archipelago, an area of rapid land uplift in the Baltic Sea, these herbivores may cause losses of up to 50% of the annual shoot production in some island populations and subpopulations of S. dioica. The genetic structures of the host populations resulting from colonization and extinction dynamics in the context of land uplift are well characterized. Island populations are differentiated and conform to expectations from genetic metapopulation models. Of potential importance to insect-plant interactions is the strong genetic differentiation occurring on a 1 m2 scale resulting from kin-structured founding and mating dynamics in the host, since this may translate into heterogeneity in host-plant quality with a potential for local selection. As a first approach to disentangling a selection dynamic, we present analyses of the consequences of herbivory for host survival, dormancy, flowering frequency and patch recruitment from a six-year time series collected from specific locations and patches on different islands.

DOES TREE SPECIES DIVERSITY AFFECT SURVIVAL OF FOREST PESTS?
Janne Riihimäki1, Pekka Kaitaniemi2, Julia Koricheva1 & Harri Vehviläinen1
1 Section of Ecology, Department of Biology, University of Turku, FIN-20014 Turku, Finland.
2 Hyytiälä Forestry Field Station, University of Helsinki, Hyytiäläntie 124, FIN-35500 Korkeakoski, Finland.

The enemies hypothesis, which states that predators and parasites are more efficient in controlling pest densities in polycultures than in monocultures, is little studied in forest ecosystems. In a two-year study, we investigated whether survival and parasitism of the autumnal moth (Epirrita autumnata) and the European pine sawfly (Neodiprion sertifer) larvae differ between young monocultures and bicultures. Survival of E. autumnata on birch was monitored in silver birch monocultures and paired mixtures with black alder, Scots pine and Norway spruce. In 2002, the larvae disappeared earlier and their final survival was lower in birch-pine mixtures than in other stand types, but in 2003 survival did not differ between stand types. Among the predators, wood ants were more abundant in birch-pine mixtures than in other stand types probably because colonies of myrmecophilic aphids were common on pines. Survival times of N. sertifer larvae were significantly shorter in pine-birch mixtures than in pure pine stands and this was also partially due to higher ant numbers in bicultures. Tree species composition did not affect larval parasitism of either species. Our results provide a partial support for the enemies hypothesis and suggest that ants are an important factor controlling pests in young forests.

SPATIAL VARIATION IN HOST PLANT QUALITY IN THE MELITAEA CINXIA METAPOPULATION

Marjo Saastamoinen & Anna-Liisa Laine
Metapopulation Research Group, Department of Biological & Environmental Sciences, PO Box 65, FI-00014, Finland.

Spatial variation in host plant quality within and among populations may be caused, for example, by genetic differences or defence metabolites induced by exploitation by other species. This variation may affect the development and survival of insects feeding on them. In the Åland Islands, SW Finland, Melitaea cinxia butterfly occurs as a classical metapopulation and the two larval host plants, Plantago lanceolata and Veronica spicata have a highly fragmented distribution. The host populations are known to vary in the concentration of iridoid glycosides and in the incidence of a powdery mildew fungus. Two separate feeding assays were conducted to see how these factors affect the larvae. The level of iridoid glycosides had no effect on larval survival. In both host plants high concentrations of iridoid glycosides were associated with increased larval growth. Additionally, results showed that in the first instar iridoid glycosides tend to retard larval development, whereas in the second and third instars they accelerate development. Larvae feeding on the fungal-infected diet developed more slowly and weighed less at the onset of diapause compared to larvae feeding on healthy host leaves. An analysis of 167 M. cinxia populations showed that over-winter survival of larval groups was 26% lower in host populations infected by the mildew than in non-infected host populations. These results suggest that spatial variation in host plant quality may affect the dynamics of the M. cinxia metapopulation.
SPATIO-TEMPORAL PATTERNS IN A HOST-ENEMY SYSTEM IN AN ARCHIPELAGO
Johan Stenberg1, Johanna Witzell2, Lars Ericson1
1 Dept. Ecology and Environmental Science, Umeå University, SE-90187 Umeå, Sweden.
2 Umeå Plant Science Centre, Dept. Forest Genetics and Plant Physiology, Swedish University of Agricultural Sciences, SE-90183 Umeå, Sweden.

We have studied the interactions between the host plant Filipendula ulmaria and its monophagous chrysomelid beetles Galerucella tenella and Altica engstroemi in a Bothnian archipelago subjected to isostatic rebound. Data for the period 1992-2003 show that G. tenella occurs throughout the archipelago experiencing outbreaks uncorrelated in time, while A. engstroemi is confined to the innermost archipelago and often show high population densities during extended periods. Data from populations, representing different time lengths of coexistence, show that the interaction between the host plant and G. tenella is shaped by reciprocal selection and adaptation, thus interpreted as a case of symmetric coevolution. However, in one part of the archipelago F. ulmaria co-occurs with Lythrum salicaria, the host of the chrysomelid G. calmariensis which shares a Hymenopteran parasitoid (Asecodes mento), with G. tenella, drastically reducing the likelihood for G. tenella to reach high densities. We conclude that interactions between host plants and their natural enemies can evolve in different directions even within limited geographical areas, and that not only spatio-temporal scales, but also associated species have to be considered.
GYPSY MOTHS GO FOR A SWIM - THE SPATIAL ECOLOGY OF THE INTERACTION BETWEEN ACENTRIA EPHEMERELLA AND SUBMERGED MACROPHYTES
Dietmar Straile, Oliver Miler & Michael Korn
Limnological Institute, Department of Biology, University of Konstanz, Germany.

The aquatic moth Acentria ephemerella causes considerable defoliation of macrophytes, which are both food and habitat for Acentria. Acentria mass outbreaks and several characteristics of its biology (e.g., female flightlessness, biased sex ratio) suggest that Acentria can be considered as an aquatic analogue to the Gypsy moth. Mesocosm experiments revealed that the absence of fish predation is probably a major determinant for mass outbreaks, especially since fish predation results into male dominance. The latter because higher predation on probably more active females. Within the Special Collaborative Program “Littoral of Lake Constance” we study several aspects of the macrophyte - Acentria – fish tritrophic system. Using remote sensing the patch dynamics and isolation of macrophytes is studied. With intensive field sampling programmes we analyse the seasonal and spatial dynamics of Acentria (and to a lesser extent of fish). These data suggest that fish predation is reduced in small and isolated patches of macrophytes. In mesocosm and aquaria experiments trophic interactions and the life history of Acentria is analysed.
NOCTUID MOTHS FEEDING ON CATTAIL (TYPHA LATIFOLIA) AS A MODEL SYSTEM FOR STUDIES ON SPATIAL ECOLOGY OF HERBIVORES
Tiit Teder & Toomas Tammaru
Institute of Zoology and Hydrobiology, University of Tartu, Vanemuise 46, EE-51014 Tartu, Estonia.

Cattails are perennial plants that form monospecific stands in damp sites. These plants are inhabited by two species of taxonomically related noctuid moths that differ in body size. Biological features of these moths allow one to easily obtain unbiased samples and density estimates. We have shown that plant vigour has a strong and predictable influence on pupal weights (fecundities) of the herbivorous moths both on temporal and spatial scales. Moreover, we documented a three_trophic link between plant vigour and parasitoid body sizes (Ecol. Entomol.  27: 94-104). We have recorded spatial indirect effects among the populations of the two herbivores: the smaller moth escapes parasitism when the larger one is abundant (Ecography 22: 79-88) while both species are equally parasitised when they occur at comparable abundances (Eur. J. Entomol. 100: 323-328). Our data base describing spatiotemporal variation in both fecundity and mortality allows one to estimate selection pressures on moth dispersal ability. We expect the selection to be considerable as these moths - unlike most other lepidopterans with non-feeding adults - have not evolved heavy abdomens impeding their mobility.
SIGNS OF SELECTIVE MOSAIC IN THE ARCHIPELAGO SEA: HOST QUALITY AND THE COMPOSITION OF HOST PLANT ASSEMBLAGE AFFECT THE HOST USE ABILITY OF THE MARINE ISOPOD IDOTEA BALTICA
Outi Vesakoski, Veijo Jormalainen, Tuija Honkanen & Jenni Rautanen
Section of Ecology, FIN-20014, University of Turku, Finland.
 
According to the geographic mosaic theory of coevolution, spatial variation in host plant quality may cause geographically varying adaptations to host use ability in the herbivore. Local adaptations may also be caused by differences in host plant assemblages. We studied 1) whether the marine isopod Idotea baltica is locally adapted to utilize sympatric bladder wrack (Fucus vesiculosus) and 2) whether the isopods are adapted to utilize the dominant host species of the local habitat (Fucus vs. Zostera marina). The isopod populations varied in their performance and in the degree of local adaptation. The quality of Fucus had an effect on isopods – in both experiments all the isopods performed best with alga originating from a northernmost locality. Moreover, the performance of the isopods was much better when fed on Fucus than on Zostera implying a universal adaptation to utilize it. Anyhow, two of the three isopod populations originating from Zostera-habitats had somewhat lower performance when feeding on Fucus and higher on Zostera than the populations from Fucus-habitats suggesting weak adaptation to local host assemblages. We conclude that spatial differences in host quality and especially in host species availability possess a weak but existing selective agent on herbivores’ host utilization efficiency.
HOST PLANT POPULATION SIZE DETERMINE CASCADING EFFECTS IN A PLANT-HERBIVORE-PARASITOID SYSTEM
Hugo von Zeipel
Department of Botany, Stockholm University, S-10691, Stockholm, Sweden.

Cascading effects from a third level consumer to the basal plant level have been documented several times in terrestrial systems. However, less is known about how these effects are influenced by characteristics of the host plant population. In the study system the perennial herb Actaea spicata is attacked by larvae of one exlcusive, specialist moth seed predator, Eupithecia immundata. The seed predator in turn is attacked by a guild of hymenopteran parasitoids. The distribution of the insects were studied among 85 populations of A. spicata. The incidence of all insect species were positively correlated to plant population size. The parasitoids were missing or rare in small to intermediate sized plant populations but abundant in large populations. However, the seed predator was missing only in very small plant populations. The effect on the plants was a unimodal relationship between population size and proportion of fruits preyed upon with higher mean levels in intermediate sized populations compared both to smaller and larger populations. The larval mortality of the seed predator, caused by parasitoids, increased with plant population size and thereby enhanced plant fitness. This was also reflected in plant population structure since the proportion of seedlings was negatively correlated to the proportion of seeds preyed upon. This study shows that multitrophic level interactions are influenced by host plant population size and that natural enemies of a specialist herbivore have long term cascading effects in terms of changes in  plant population structure.

Back to Plant-Insect Interaction Workshop Page

Back to Spatial Ecology home page