Host plant variation in mortality of the leaf-folding sawfly on the arroyo willow

Abstract. 1. We tested the hypothesis that survival and sources of mortality of the leaf-folding sawfly ( Phyllocolpa sp.) varied significantly among host plants of the arroyo willow ( Salix lasiolepis Bent ham).
2. Survival of the leaf folder differed among field and potted willows in a common environment in two of three cases, and sources of mortality differed among plants in four of five cases.
3. Egg mortality differed among field plants but not among the potted willow plants.
4. Larval mortality and parasitism differed among field and potted willows in 2 years, and appeared to be compensatory mortality sources.
5. Leaf folder density among plants was not generally correlated with percentage of leaf folds with no egg (galls formed but no subsequent oviposition), percentage survival, or percentage mortality, indicating a general lack of density dependence.
6. The proportion of folds with no egg oviposited differed significantly among field and potted plants, and was only correlated with survival or sources of mortality in one of three years.
7. The data support the hypothesis that host plant genotype affects the interaction of the leaf folder with its natural enemies, and thus represents a three trophic level interaction.     

Genetic variation in resistance and tolerance to insect herbivory in Salix cordata

1. This study investigated whether sand-dune willow Salix cordata , exhibits genetic variation in resistance and tolerance to herbivory.
2. A field experiment using cuttings from nine willow clones demonstrated genetic variation in resistance to the specialist herbivore Altica subplicata , as measured by beetle densities. Willow clones differed significantly in both total biomass and leaf trichome densities, and herbivore densities were marginally correlated with both of these parameters.
3. Tolerance to herbivory was measured in a greenhouse experiment by comparing growth response of plants experiencing 50% artificial defoliation and plants experiencing no defoliation. Clones showed significant differences in tolerance to herbivory for some growth measures (changes in height and number of leaves), but not for other growth measures (stem diameter growth and final biomass).
4. Despite the significant genetic variation in both resistance and tolerance, no trade-off was found between resistance and tolerance to herbivory.     

Direct and indirect effects of plant genetic variation on enemy impact

Abstract.

• 1 The Tritrophic and Enemy Impact concepts predict that natural enemy impact varies: (a) among plant genotypes and (b) may depend on the abundance of heterospecific herbivores, respectively. I tested these predictions using three herbivore species on potted, cloned genotypes of Salik sericea Marshall in a common garden experiment.
• 2 Densities of the leaf miner (Phyllonorycter salicifoliella (Chambers)) and two leaf galling sawflies (Phyllocolpa nigrita (Marlatt) and Phyllocolpa eleanorae Smith and Fritz) varied significantly among willow clones, indicating genetic variation in resistance.
• 3 Survival and natural enemy impact caused by egg and larval parasitoids and/or unknown predators differed significantly among willow clones for each of the three herbivore species, indicating genetic variation in survival and enemy impact.
• 4 Survival of Phyllonorycter was negatively density-dependent among clones.
• 5 Survival of Phyllonorycter and Phyllocolpa eleanorae were positively correlated with densities of heterospecific herbivores among clones and parasitism of these species were negatively correlated with densities of the same heterospecific herbivores among clones.
• 6 At least for Phyllonorycter this positive correlation may suggest either facilitation of survival between herbivore species, which do not share natural enemies, or an apparent interaction caused by host plant genetic variation.
• 7 Among clones, egg parasitism of Phyllocolpa eleanorae was weakly positively correlated with density of Phyllocolpa nigrita. Since these species share the same Trichogramma egg parasitoid, this interaction could support the hypothesis of apparent competition.

     

Host plant variation in plant-mediated indirect effects: moth boring-induced susceptibility of willows to a specialist leaf beetle

Abstract.  1. We examined the plant-mediated indirect effects of the stem-boring moth Endoclita excrescens (Lepidoptera: Hepialidae) on the leaf beetle Plagiodera versicolora (Coleoptera: Chrysomelidae) in three willow species, Salix gilgiana , S. eriocarpa , and S. serissaefolia.
2. When the stem-boring moth larvae damaged stems in the previous year, willows were stimulated to produce vigorously growing lateral shoots on these stems. These new lateral shoots were significantly longer and the upper leaves had significantly higher nitrogen and water content than current-year shoots on unbored stems, although the carbon content and leaf dry mass were not different between lateral and current-year shoots.
3. In the field, leaf beetle larvae and adults had significantly greater densities on lateral shoots of bored stems than on current-year shoots of unbored stems. A laboratory experiment showed that female beetles had significantly greater mass and fecundity when fed on leaves of newly-emerged lateral shoots. Thus, the stem-boring moth had a positive effect on the temporally and spatially separated leaf beetle by increasing resource availability by inducing compensatory regrowth.
4. The strength of the indirect effects on the density and performance of the leaf beetle differed among willow species, because there was interspecific variation in host quality and herbivore-induced changes in plant traits. In particular, we suggest that the differences in magnitude of the changes among willow species in shoot length and leaf nitrogen content greatly affected the strength of the plant-regrowth mediated indirect effect, coupled with host-plant preference of the leaf beetle.     

Variation in herbivore density among host plants and its consequences for community structure

Summary The densities of four species of gall-forming sawflies were found to vary significantly among willow host plant clones. Two of the speices varied among host plants at four sites in each of three years. The other two species varied in density among host plants at most of the sites in two of the three years. Total sawfly density also varied significantly among clones. Individual species densities on willow clones were significantly positively correlated between years when all sites were combined and frequently when sites were considered separately. Most pairwise species combinations were independent in density between years, but some negative correlations existed between the stem galler and the leaf galler. Gall-former densities also were largely independent among clones within years with all sites combined and with sites considered separately. The significant correlations were nearly all positive. At all four sites the combination of significant variation in sawfly densities among willow clones in the field and independence of species densities among clones resulted in significantly different communities (relative abundance of species) among willow clones in three years. Although sawfly abundances differed substantially among the four sites, this remained true. It is argued that the pattern of community structure among clones is the result of variation in host plant quality of clones. We propose an hypothesis to account for patterns of herbivore species associations based on intrapopulation host plant variation.     

Preference and performance of a gall-inducing sawfly: a test of the plant vigor hypothesis

To test the Plant Vigor Hypothesis, we determined female oviposition preference of Phyllocolpa leavitii (Hymenoptera: Tenthredinidae) for shoot lengths on 20 clones of Salix discolor (Salicaceae) and examined larval performance by determining larval survival based on clone, shoot length, and leaf length for each Phyllocolpa gall observed. Sawfly galls were found on significantly longer shoots compared to shoots without galls, and shoots with 2, 3, 4, and 5 galls were successively longer than shoots with fewer galls. Furthermore, a much higher proportion of longer shoots had galls that did shorter shoots. These data demonstrate strong sawfly oviposition preference for long shoots. However, when shoot length was adjusted for numbers of available leaves per shoot, the number of galls per available leaf showed a weaker, but still positive relationship with shoot size. Using a logistic regression on survival of Phyllocolpa larvae, we detected highly significant effects of clone, shoot length, leaf length, and the square of leaf length on survival. Additionally, much higher sawfly survival on the long shoots of one clone caused a significant clone by shoot-length interaction. Survival was positively correlated with shoot length, indicating directional selection favoring oviposition on longer shoots; still, shoot length explained only 11.8% of the variation in survival. The directional component of selection favored oviposition on smaller leaves as survival was inversely related to leaf length; however, disruptive selection for leaf length oviposition preference was also detected, with survival of Phyllocolpa galls lowest on intermediate leaf lengths. This study provided evidence supporting the preference prediction of the Plant Vigor Hypothesis. However, much of the data contradicted the performance prediction of the Plant Vigor Hypothesis, with only a modest amount of evidence supporting the performance prediction.     

Variable responses of natural enemies to Salix triandra phenotypes with different secondary chemistry

Plant phenotypes often differ in their resistance to natural enemies, but the mechanism for this has seldom been identified. The aim of this study was to determine if the spatial patterns of phenotype use of a highly specialized insect herbivore (the galling sawfly Pontania triandrae ) in a natural willow population can be related to phenotypic variation in plant secondary chemistry. Furthermore, we tested if traits that confer resistance to one type of natural enemy, i.e. the galling sawfly, also confer resistance to others, in our case a leaf beetle Gonioctena linnaeana and the rust fungus Melampsora amygdalinae . We identified 18 phenotypes with high and 18 phenotypes with low gall density in our field population and determined gall densities, the degree of leaf damage and rust infection on each phenotype and collected leaves for chemical analyses. The concentration of phenolics was higher in phenotypes with high density of galls suggesting that this galling sawfly may use phenolics as oviposition cues. Rust infection showed the opposite pattern, with lower levels on clones with high concentration of phenolics, while leaf damage by G. linnaeana did not differ between clone types. This indicates that these important natural enemies may assert divergent selection on willow phenotypes and that this might provide a mechanism for maintaining phenotypic variation within willow populations.     

Spatial Distribution of Galls Caused by <Emphasis Type="Italic">Aculus tetanothrix</Emphasis> (Acari: Eriophyoidea) on Arctic Willows

The distribution of galls caused by Aculus tetanothrix (Acari: Eriophyoidea) on three Salix species was studied. The factors influencing this distribution were analysed, i.e. willow species, study area and shoot length. Spatial pattern of gall distribution within the shoot was also examined. The study was conducted in Russia, Kola Peninsula. Densities of galls caused by A. tetanothrix differed significantly among willow species. Considerably higher gall density was recorded in the White Sea coast than in the Khibiny Mountains. This may be explained by the influence of a milder maritime climate that favors mite occurrence compared to a harsh and variable mountain climate that limits mite abundance. There was no relationship between the gall density and the shoot length. The highest density of galls was recorded on the inner offshoots; within the offshoot, there was a maximum density on the fifth leaf. This pattern was repeatable for all shoots studied, independent of the study area, willow species and length of shoots, suggesting the optimal conditions for A. tetanothrix exist on leaves in the middle part of a shoot. This distribution pattern may be an effect of the trade-off between the costs and benefits resulting from leaf quality and mite movement along the shoot. This hypothesis, however, needs to be tested experimentally.     

Growth of coyote willow and the attack and survival of a mid-rib galling sawfly,Euura sp.

We studied the relationship between variation in age and shoot characteristics of the host plant Salix exigua Nuttall (coyote or sandbar willow) and the attack and survival of Euura sp. (an unnamed leaf-midrib galling sawfly). Variation in shoot characteristics resulted from reduced growth as willow ramets aged. Mean shoot length per ramet and mean longest leaf length per shoot decreased by 95% and 50% respectively between 1- and 9-year-old willow ramets. All measured shoot characteristics-shoot length, longest leaf length, number of leaves per shoot, and mean internode length-were significantly negatively correlated with ramet age (r ² ranged from –0.23 to –0.41). Correlations between shoot characteristics were highly positive, indicating that plants also grew in a strongly integrated fashion (r ² ranged from 0.54 to 0.85). Four hypotheses were examined to explain sawfly attack patterns. The host-plant hypothesis was supported in explaining enhanced larval sawfly survival through reduced plant resistance. As willow ramets aged, the probability of Euura sp. attack decreased over 10-fold, from 0.315 on 1-year-old ramets to 0.024 on 2- to 9-year-old ramets. As shoot length increased, the probability of sawfly attack increased over 100-fold, from 0.007 on shoots <100 mm, to 0.800 on shoots in the 1001–1100 mm shoot length class. These attack patterns occurred even though 1-year-old ramets and shoots >500 mm each represented less than 2% of the total shoots available for oviposition. Host plant induced mortality of the egg/early instar stage decreased by 50% on longer leaves and was the most important factor determining survival differences between vigorous and non-vigorous hosts. Sawfly attack was not determined by the resource distribution hypothesis. Although shoots <200 mm contained 82% of the total leaves available, they contained only 43% of the galls initiated. The attack pattern also was not explained by the gall volume hypothesis. Although gall volume increased on longer shoots, there was no significant variation in mid or late instar mortality over shoot length, as would be expected if food resources within smaller galls were limited. The natural enemy attack hypothesis could not explain the pattern of oviposition since predation was greater on longer shoots and leaves. In addition, larval survival was related to oviposition behavior. Due to a 69% reduction in late instar death and an 83% reduction in parasitism, survival of progeny in galls initiated close to the petiole base was 2.8 times greater than in galls initiated near the leaf tip. A 75% reduction in gall volume over this range of gall positions may account for the observed increases in late instar mortality and parasitism.     

Patterns of intra- and interspecific association of gall-forming sawflies in relation to shoot size on their willow host plant

Summary Four species of gall-forming sawflies were each frequently found to have clumped distributions among shoots within their willow host plant across four sites and three years. When all species were considered together by clone, year, and site, species showed independence of distribution among shoots two thirds of the time and showed positive covariance one third of the time. When pairs of species were considered separately, but clones were combined within sites and years, 60% of the chi-square tests of association were significant. All but one of the significant tests showed positive associations between pairs of species. The stem galler was positively associated with the leaf folder at all sites in all years, and the petiole galler was positively associated with the stem galler and leaf folder for most year by site combinations. When species paris co-occurred on shoots they were usually found at the same or higher density as when found alone on shoots. Only 2 of 100 tests showed a depressed density of a species when co-occurring on shoots with heterospecifics.All sawfly species were found on shoots that were significantly larger (mean node number) than on shoots without sawflies, and species responded to shoot size variation similarly. Sizes of shoots occupied by heterospecific species combinations were usually significantly larger than shoots with only conspecifics, for all species. These data supported the hypothesis that similar species' responses to within-plant variation would lead to positive rather than negative or random species associations. The data do not support the hypothesis that interspecific competition was important in determining shoot choice or species density.     

Shoot growth in willow (Salix viminalis) in relation to abscisic acid, plant water status and photoperiod

The decline in growth rate of field-grown willow trees in Aberystwyth, U.K., began in mid-summer and was followed by the senescence and abortion of shoot tips. These events were not triggered by a decline in the length of the natural photoperiod but were coincident with low leaf water potentials that developed in summer. Transient increases in the abscisic acid (ABA) content of shoot tips were observed during the period of declining water potential. These increases were roughly coincident with the onset of growth decline and preceded abortion and senescence of shoot tips. Under controlled conditions growth of both rooted cuttings and potted plants was arrested by short days (8 h) without any increase in tip ABA levels. Growth of rooted cuttings under long days (16 h) was inhibited by exogenous ABA; this inhibition could be relieved by addition of gibberellic acid (GA3) to the nutrient solution. Growth of aseptically cultured apices was also inhibited by ABA; this inhibition was relieved by joint application of GA9 and zeatin riboside.     

Oviposition choice and larval survival of Dasineura marginemtorquens (Diptera: Cecidomyiidae) on resistant and susceptible Salix viminalis

Abstract.

• 1 Substantial intraspecific variation exists in Salix viminalis resistance to the gall midge Dasineura marginemtorquens. Earlier work has found this variation to have a large genetic component. Willow clones are stable in their resistances between midge generations and different nutrient levels in both field and laboratory culture.
• 2 This study reports the results of laboratory experiments on female oviposition choice and larval survival on potted plants from clones that are very different in resistance as determined in field studies.
• 3 In choice experiments using pairs of plants, the average female midge did not prefer susceptible willow clones over resistant ones for oviposition. In about one third of the replicates, midges actually laid more eggs on the resistant clone. Further work is necessary to examine the nature of variation among midges in discrimination of these plant types.
• 4 Resistance is manifested as great differences in larval survival. Six days after oviposition survival was 92% on susceptible plants but only 6% on resistant ones. Galls developed on all of the susceptible plants, while in 73% of the resistant plants galls were not even initiated.
• 5 The plant traits causing resistance are enigmatic. Larval behaviour suggests that resistant plants interfere with feeding behaviour. On resistant plants, most larvae wander for more than 24 h without initiating any galls before dying. On susceptible plants many first instar larvae begin feeding and initiate galls within this period.

     

Rapid gall midge adaptation to a resistant willow genotype

Abstract 1 We conducted two experiments to investigate why a basket willow Salix viminalis L. genotype, known to be highly resistant to the leaf-roller gall midge Dasineura marginemtorquens (Bremi), should support very high gall densities in a field plantation at Tälle, south Sweden.
2 The first experiment was a field test of the hypothesis of fine-scale host adaptation in the gall midge/willow system. Support for the hypothesis would be established if midges originating from resistant willows and those originating from nearby susceptible willows differed in their abilities to initiate galls and complete development on resistant plants.
3 The objective of the second experiment was to explore whether there was a genetic basis to the trait for virulence in the midge population and to investigate any potential trade-offs this trait may entail.
4 Our results indicate that there was a fine-scaled microgeographic genetic structure to the midge population at Tälle. Midges originating from resistant plants had a heritable trait that enabled them to establish galls on resistant plants.
5 Midges able to initiate galls on the resistant genotype had longer developmental time on the susceptible genotype. This suggests that there is a physiological cost associated with being adapted to the resistant willow genotype.
6 We suggest that driving forces behind the observed host adaptation are selection imposed on the midge population by very strong willow resistance and restricted gene flow in the midge populations due to the special life history features of D. marginemtorquens .     

Plant phenology-mediated indirect effects: The gall midge opens the phenological window wider for a leaf beetle

We examined whether larvae of the gall midge Rabdophaga rigidae (Diptera: Cecidomyiidae) can modify the seasonal dynamics of the density of a leaf beetle, Plagiodera versicolora (Coleoptera: Chrysomelidae), by modifying the leaf flushing phenology of its host willow species, Salix serissaefolia and Salix eriocarpa (Salicaceae). To test this, we conducted field observations and a laboratory experiment. The field observations demonstrated that the leaf flushing phenology of the willows and the seasonal dynamics of the beetle density differed between shoots with stem galls and shoots without them. On galled shoots of both willow species, secondary shoot growth and secondary leaf production were promoted; consequently, leaf production showed a bimodal pattern and leaf production periods were 1 to 2 months longer than on non‐galled shoots. The adult beetle density on galled shoots was thus enhanced late in the season, and was found to change seasonally, synchronizing with the production of new leaves on the host willow species. From the results of our laboratory experiment, we attributed this synchrony between adult beetle density and willow leaf flush to beetles’ preference to eat new leaves rather than old. Indeed, beetles consumed five times more of the young leaves when they were fed both young and old leaves. These results indicate that stem galls indirectly enhance the adult beetle density by enhancing food quality and quantity late in the beetle‐feeding season. We therefore conclude that midge galls widen the phenological window for leaf beetles by extending the willows’ leaf flush periods.     

Gall distribution as a compromise between the optimal gall-site selection and the synchrony to host-plant phenology in the aphid Kaltenbachiella japonica

Gall-site selection by the aphid Kaltenbachiella japonica was evaluated in relation to leaf position in a shoot, and gall positions within a leaf. First-instar fundatrices induce closed galls on the midribs of host leaves, and several galls were often induced on one leaf. Leaves with many galls were often withered before emergence of sexuparae from the galls. Within a leaf, gall volume was positively correlated with the sum of lateral-vein length in the leaf segment at which the gall was induced. The observed pattern in gall volume among the leaf segments corresponded with that in the lateral-vein length. These results show that a foundatrix selects the most vigorous position within a leaf to produce more offspring. Although distal leaves grew faster than did basal leaves, gall density was highest on leaves at the middle order when a shoot has more than seven leaves. Optimal gall-site selection seems to be constrained by the asynchrony in timing between the hatching of fundatrices and leaf growth within a shoot. These results suggest that the observed gall distribution is affected by both the distribution of suitable galling sites within a leaf and the synchrony with leaf phenology of the host plant.     

Does nitrogen fertilization have an impact on the trade-off between willow growth and defensive secondary metabolism?

The effect of nitrogen fertilization on the phytomass production, shoot length and leaf secondary phenolics in nine Salix myrsinifolia clones was investigated. Cuttings taken from 1-year-old and 2-year-old shoot parts of field cultivated clones were grown at three concentrations of nitrogen (7, 150 and 300 ppm) in a greenhouse for one growing season. The willow clones differed significantly in phytomass yield and secondary phenolics content. Nitrogen fertilization affected significantly the growth and secondary metabolism of willow clones. In most clones, the addition of nitrogen from a sub-optimum concentration (7 ppm) to an optimum concentration (150 ppm) appeared to reduce the amounts of salicortin, chlorogenic acid and unknown salicylate and increased shoot phytomass, but a supraoptimum nitrogen concentration (300 ppm) resulted in highly variable growth and secondary phenolic responses. A significantly negative correlation between leaf phytomass and amount of total phenolics at sub-optimum and optimum N-treatments indicates trade-off between growth and secondary metabolism in willow clones at these treatments. However, the leaf phytomass:total amount of phenolics ratio varied significantly among clones, and in all clones it was not significantly lower at sub-optimum N-treatment than at optimum N-treatment.     

The effects of host-plant properties on gall density,gall weight and clone size in the aphidAploneura lentisci (Pass.) (Aphididae: Fordinae) in Israel

Summary Samples of shoots ofPistacia lentiscus carrying galls of the aphid,Aploneura lentisci, were collected at three localities in Israel. Shoots growing near pruning scars carried more galls than elsewhere on the plant, but these galls weighed less and contained fewer aphids (smaller clones). The proportion of empty galls increased with gall density. Crowding of galls at such sites may be due to the early burst of buds at the time of aphid emergence from the overwintering eggs, and not to active search for preferred sites. Shoots bearing larger numbers of leaves carried heavier galls, which contained larger aphid clones. The position of the galled leaf on the shoot had no effect on gall weight nor on clone size. The physiological condition of the plant may be an important environmental (ecological) factor affecting the variation in clone-size and in aphid morphology among galls.     

Consequences of host plant chemical and physical variability to an associated herbivore

The effects of watering and fertilizer treatments on the vigor and biochemistry of the willow,Salix lasiolepis, and subsequent colonization and survivorship of its gallforming herbivore,Euura lasiolepsis, were investigated in two field experiments. Some plants received low (LW), intermediate (MW) or high (HW) levels of water as treatments, while others received no (OF), low (LF) or high (HF) fertilizer levels. In the watering experiment, plant protein concentrations decreased, while growth rate and number of galls per plant increased with increased water treatments. Plant growth proved to be the best correlate of sawfly attack. Sawfly survivorship increased slightly with greater watering, and phenol concentrations showed no pattern among treatments. In the fertilization experiment, leaf protein increased with fertilization, although shoot length, number of galls and survivorship ofE. lasiolepis survivorship were greatest in intermediate treatment plants. In both experiments, plant growth, rather than protein or phenol levels, was the best predictor of sawfly attack and survivorship. In a natural experiment with galls on wild plants, galled tissue had significantly greater protein concentrations and lower phenol concentrations than did ungalled tissue. We suggest that gallformers modify host plant biochemistry within willow galls, which may explain why the chemical parameters of ambient plant quality we tested were less predictive than plant growth.     

Above-ground competition does not alter biomass allocated to roots in Abutilon theophrasti

We tested whether plants allocate proportionately less biomass to roots in response to above-ground competition as predicted by optimal partitioning theory. Two population densities of Abutilon theophrasti were achieved by planting one individual per pot and varying spacing among pots so that plants in the two densities experienced the same soil volume but different degrees of canopy overlap. Density did not affect root:shoot ratio, the partitioning of biomass between fine roots and storage roots, fine root length, or root specific length. Plants growing in high density exhibited typical above-ground responses to neighbours, having higher ratios of stem to leaf biomass and greater leaf specific area than those growing in low density. Total root biomass and shoot biomass were highly correlated. However, storage root biomass was more strongly correlated with shoot biomass than was fine-root biomass. Fine-root length was correlated with above-ground biomass only for the small subcanopy plants in crowded populations. Because leaf surface area increased with biomass, the ratio between absorptive root surface area and transpirational leaf surface area declined with plant size, a relationship that could make larger plants more susceptible to drought. We conclude that A. theophrasti does not reallocate biomass from roots to shoots in response to above-ground competition even though much root biomass is apparently involved in storage and not in resource acquisition.     

Preference–performance relationship in the gall midge Rabdophaga rosaria: insights from a common‐garden experiment with nine willow clones

1. Oviposition preferences of herbivorous insects are predicted to match offspring performance on different host taxa or on conspecific host genotypes. In gall‐inducing insects, host‐plant properties such as growth rate and gall size, which are determined by plant genotype and growing conditions, may have a significant impact on offspring performance and, hence, should influence oviposition site selection. 2. The present study investigated host preference of the European rosette willow gall midge Rabdophaga rosaria (Loew) in relation to offspring success on seven clones of Salix myrsinifolia Salisb. and two naturally hybridised S. myrsinifolia × phylicifolia L. clones growing in a replicated design in an experimental field under two fertilisation regimes. For each clone, the average growth rate, number of shoot tips, and leaf and gall size were determined, and their effects on midge preference and larval survival were examined. 3. Main shoot height, number of shoot tips, and gall size were significantly affected by clone. The midges clearly preferred certain clones over the others, but preferences were not related to willow growth traits or to gall size. Survival probability was higher in large than in small galls, but females did not prefer large‐leaved clones that produced the biggest rosette galls. Midge oviposition was also uncorrelated with prior rates of leaf‐rust infection and with feeding preferences of voles and folivorous insects. 4. The weak preference–performance relationship of R. rosaria within S. myrsinifolia is probably explained by evolutionary constraints that prevent generalist insects from achieving an ability to discriminate among conspecific hosts of variable quality.