Oct 27, 1997 - experimental studies in which insect performance was measured on woody plants subjected to water stress, pollution, and/or shading. Overall ...
P1: ARS
October 27, 1997
16:59
Annual Reviews
AR048-10
Annu. Rev. Entomol. 1998. 43:195–216 c 1998 by Annual Reviews Inc. All rights reserved Copyright
INSECT PERFORMANCE ON EXPERIMENTALLY STRESSED WOODY PLANTS: A Meta-Analysis Julia Koricheva and Stig Larsson Department of Entomology, Swedish University of Agricultural Sciences, P.O. Box 7044, S-750 07 Uppsala, Sweden
Erkki Haukioja Laboratory of Ecological Zoology, Department of Biology, University of Turku, FIN-20014 Turku, Finland KEY WORDS:
plant stress, insect feeding guild, water stress, pollution, shading
ABSTRACT In this review, we test the hypothesis that abiotic stress increases the suitability of plants as food for herbivores. We conducted a meta-analysis that included 70 experimental studies in which insect performance was measured on woody plants subjected to water stress, pollution, and/or shading. Overall, plant stress had no significant effect on insect growth rate, fecundity, survival, or colonization density. We found great variation, however, in the magnitude and direction of insect responses among studies, most of which was related to insect feeding guild. In general, boring and sucking insects performed better on stressed plants, whereas plant stress adversely affected gall-makers and chewing insects. Reduction in performance of chewers was greater on stressed slow-growing plants than on stressed fast growers. Reproductive potential of sucking insects was increased by pollution but reduced by water stress. In some cases where sample sizes were small or the treatment periods short, apparent differences in insect responses to stress were probably artifacts due to inappropriate experimental design.
OVERVIEW AND SCOPE Forest entomologists have long recognized that insect outbreaks are often preceded by periods of atypical weather conditions, especially drought, excessive precipitation, or unusually hot or cold temperatures (reviewed in 65, 74, 75). 195 0066-4170/98/0101-0195$08.00
P1: ARS
October 27, 1997
196
16:59
Annual Reviews
AR048-10
KORICHEVA ET AL
Similarly, in recent years insect outbreaks have often been reported from areas exposed to air pollution (42). It has been suggested that plants subjected to such stressful conditions become more susceptible to herbivorous insects owing to the plant’s increased nutritional quality and/or reduced concentrations of defensive chemicals (100, 116, 121–123). Although this plant stress hypothesis (PSH) has achieved paradigm status, most evidence supporting it is circumstantial, consisting of numerous observations linking insect abundance to climatic factors (74, 75). Correlations between stressful conditions and outbreaks do not necessarily mean that changes in host plant quality are responsible for triggering insect outbreaks. Increases in insect densities can be due to direct effects of the environment on the insect as well as to a decrease in the abundance of its natural enemies. Thus, to obtain unambiguous results concerning the role of host plant stress in triggering insect outbreaks, experiments must be carried out in a controlled environment. Such experiments, however, have provided conflicting results—insect performance has been reported to increase, decrease, or remain unchanged in response to plant stress (63). Several recent reviews have questioned the generality of the PSH, and attempts have been made to modify the hypothesis in order to take into account variation among insect types, host plants, and experimental design (52, 63, 85, 86, 95, 114, 119). Most reviews have taken a traditional narrative approach, providing testable predictions rather than statistical analyses of the importance of different factors affecting insect responses. Two quantitative reviews (114, 119) compared the proportion of documented cases of positive insect responses to plant stress with the proportion of negative responses. However, such a “votecounting” approach does not allow the magnitude of stress effects on insects to be estimated and therefore cannot address the question of whether plant stress effects on insect performance are strong enough to affect insect population dynamics. To reach consensus as to the generality and predictive power of the PSH, a formal set of statistical procedures should be applied when assessing the data generated in stress experiments, rather than rely on individual interpretations. Meta-analysis (MA) is a statistical method that can be used to combine results from independent experiments (2, 31). Unlike “vote-counting,” MA allows assessment of the magnitude of the effects (effect sizes) across studies. MA can also be used to estimate whether the overall effect is significantly different from zero and to test whether characteristics of the study organisms, or of the experimental design, influenced the magnitude of the observed effect. Thus, MA is less subjective than narrative reviews, which makes it especially useful for resolving conflicts in areas where the results are variable across studies, as in the case with the tests of the PSH.
P1: ARS
October 27, 1997
16:59
Annual Reviews
AR048-10
PLANT STRESS AND INSECT PERFORMANCE
197
In this review, we use MA in order to (a) estimate the magnitude of the overall effect of plant stress on insect performance and (b) assess the importance of insect type, host-plant type, nature of the stressing agent, and differences in experimental design in shaping insect responses to plant stress. The term plant stress is used here to refer to any environmental factor potentially unfavorable to a plant (67). Our focus is on insect responses to relatively short-term stress episodes (shorter than the life span of host plants).
THE DATABASE Our database included published experimental studies in which insect responses to stressed woody plants have been investigated. Our primary bibliographic sources consisted of recently published reviews and monographs on insect responses to plant stress (42, 114, 124). The reference list was updated by conducting key-word searches using Current Contents (Agriculture, Biology, and Environmental Sciences) and Biological Abstracts. Each study had to meet three criteria for inclusion in the analysis. 1. Plants had to have been stressed experimentally, and the treatment had to have had an appropriate control. In the case of pollution stress, we also included studies in which insects of known origin were reared on plants growing naturally or planted along a pollution gradient from a pollution source. 2. At least one of the following insect response parameters had to have been measured: relative growth rate (RGR, used for insects only in this review), developmental time, pupal or adult weight, potential or realized fecundity, survival, colonization density. 3. For each parameter measured, means, some measure of variance (standard deviation, standard error, confidence intervals), and the sample sizes of control and experimental groups had to have been reported in the article as numerical or graphical data, or they had to have been available from the authors. A total of 70 studies on drought and waterlogging stress (n = 36), pollution (n = 31), and shading (n = 7), alone or in combination, published during 1958– 1996, met our criteria for inclusion. The studies covered a wide diversity of taxa (66 insect species and 40 species of plants). We classified insects into one of five feeding guilds: sucking, chewing, mining, galling, and boring insects. The first two guilds included only freeliving insects, while the last three contained internal feeders. With respect to their degree of host specialization, insects were categorized as monophagous (feeding on one genus of plants) or polyphagous (feeding on more than one genus of plants). Insects were also classified as flush feeders or senescence feeders depending on the timing of their feeding period and the type of plant tissue on which they preferred to feed (124). When primary studies lacked the necessary information for the above classification, an additional literature
P1: ARS
October 27, 1997
198
16:59
Annual Reviews
AR048-10
KORICHEVA ET AL
search was conducted. Plant growth rates were scored from 1 (very slow) to 5 (very fast) (70). Plant ontogenetic stage was categorized as seedling (1–5 years old), sapling (from 5 years of age to first reproductive event), or mature. Experiments were classified as field, laboratory, field cage, or field-laboratory (when stress treatments were conducted in the field while insect bioassays were conducted in the laboratory). Various insect performance parameters can be differentially affected by plant stress (52). Therefore, insect responses, measured in terms of RGR, fecundity, survival, and colonization density, were analyzed in four separate MAs. Signs of responses measured in terms of developmental time of larvae were reversed in order to convert them to RGR responses (i.e. decrease in developmental time was regarded as a positive effect that leads to increased RGR). Female body size, pupal weight, and potential fecundity were combined into a single MA because a linear relationship has often been reported between these parameters (e.g. 23, 39, 79, 85; but see 10, 66). In this class, we also included studies on sucking insects in which one or several females were introduced on a plant and allowed to reproduce for a certain period, after which the progeny were counted. We assumed that these experiments assessed the reproductive potential of sucking insects, although differential mortality of progeny might have been a confounding factor in some cases. The category of studies that measured colonization responses included field studies in which plants were either artificially infested with insects by the experimenter or were naturally colonized by insects. Colonization density reflects mainly feeding and/or oviposition preference, but it sometimes also reflects survival. In some cases, results for several insect or plant species were reported in the same study, or plants were subjected to different treatments within a study. Although different experiments reported in the same publication are not necessarily independent of one another, the loss of information caused by omission of such non-independent comparisons might bias the results even more than their inclusion. Therefore, we included such data into the analysis (151 comparisons from 70 studies) but used a conservative significance level (P < 0.01) to assess the effects of different factors on the outcome of the experiment, as recommended by Gurevitch et al (32). When results were presented for several sampling dates, we selected the date of the largest difference between treatment and control. When more than one level of a given type of stress was applied, the treatment resulting in the greatest difference in insect performance from the control was used in the analysis.
META-ANALYSIS All analyses were carried out using the MetaWin statistical program (101). For each comparison, an estimate of the magnitude of the treatment effect (effect
P1: ARS
October 27, 1997
16:59
Annual Reviews
AR048-10
PLANT STRESS AND INSECT PERFORMANCE
199
size, d ) was calculated as the difference between the means of the experimental and control groups divided by the pooled standard deviation and weighted by a correction term that removes small–sample-size bias (31). In cases where means and measures of variance had to be determined from a graph, figures were enlarged and manually digitized. Effect sizes were combined across studies by using the mixed effects model of MA (31) to obtain the grand mean effect size (d++), an estimate of the overall effect of plant stress on insect performance parameters. Plant stress was considered to have a statistically significant effect on a given insect response variable if the 95% confidence interval (CI) of d++ did not overlap zero. The absolute magnitude of d was considered large when it was greater than 0.8, moderate when it equaled 0.5, and small when it was less than 0.2 (16). To test the effects of different sources of variation on the magnitude and direction of insect response, studies were subdivided on the basis of the types of insect, host plant, stress, and experiment. The mean effect size (d+) was then calculated for each class. The between-class homogeneity (QB) was calculated and tested against a Chi-square distribution to determine if classes differed significantly from one another. This test may be flawed when the assumptions of normality are not met, as is often the case for variables expressed as percentages. Therefore, for survival data we conducted randomization tests and used biascorrected percentile bootstrap confidence intervals for d+ (1). Because our analysis is based only on published studies, effects of plant stress on insect performance could be overestimated if statistically significant results, and thus large values of d, are more likely to be published (the “file drawer problem”; 21, 102). This possibility was assessed by calculating the fail-safe sample size, i.e. the number of unpublished studies showing no treatment effect that has to be added to the analysis in order to reduce the mean effect size (d+) to a minimum meaningful value, d = 0.2 (102). Large values of the fail-safe sample size indicate that the results of the MA are robust with respect to the “file drawer problem.”
SOURCES OF VARIATION IN INSECT RESPONSES TO PLANT STRESS Overall effects of plant stress on insect performance were small and nonsignificant, contrary to the prediction of the PSH. However, values of d varied considerably among studies, ranging from strongly negative to highly positive (Figure 1).
Variation Related to Insect Traits FEEDING GUILD The consequences of stress-induced changes in host plants on insects can vary depending on their modes of feeding (feeding guilds). Larsson (63) tentatively ranked insect guilds in terms of the degree to which
P1: ARS
October 27, 1997
200
16:59
Annual Reviews
AR048-10
KORICHEVA ET AL Table 1 Effects of plant stress on performance of different insect feeding guilds Performance parameter
Guild
Number of Comparisons
RGR
Chewers Suckers
Fecundity
95%CIb d+a
Lower
Upper
28 10 (20)c
−0.09 −0.31 0.59 0.24
0.12 0.95
Chewers Suckers Miners
33 (60) 25 (23) 2
−0.56 −0.82 0.38 0.06 0.42 −0.65
−0.30 0.69 1.49
Survival
Chewers Suckers Miners Gallers
9 3 2 2 (25)
−0.27 0.14 0.78 −2.70
−1.00 −0.58 −0.04 −5.44
0.45 1.46 1.76 −0.09
Colonization
Chewers Suckers Miners Gallers Borers
3 7 9 5 (31) 11 (47)
−0.15 −0.31 0.35 −1.43 1.05
−1.12 −1.00 −0.29 −2.21 0.45
0.82 0.38 0.98 −0.65 1.64
aMean effect size. The absolute magnitude of d was considered + large when it was greater than 0.8, moderate when it equaled 0.5, and small when it was less than 0.2 (16). b95% confidence interval (CI). The effect of plant stress was considered statistically significant if the 95% CI did not overlap zero. cValues in parentheses are fail-safe sample sizes that show the number of additional studies with small treatment effects required to lower the significant d+ to a minimum meaningful value (0.2).
they benefit from plant stress. Cambium feeders were considered to benefit most and received the highest ranking, followed by sucking insects, mining insects, chewing insects, and gall-forming insects (which were adversely affected). In our analysis, the effect of plant stress differed significantly (P 0.01) among insect feeding guilds for all measures of insect performance except survival (P = 0.229), and the results were generally consistent with the above ranking (Table 1, Figure 1). The RGR and reproductive potential of sucking insects increased on stressed plants, whereas survival and colonization were not significantly affected. The fecundity of chewing insects was decreased by plant stress, but their RGR, −−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−→ Figure 1 Effect sizes and variance of individual studies on insect responses to plant stresses included in the meta-analysis (MA). Each bar represents a single comparison. Grand mean effect sizes (d++), 95% confidence intervals (CI), and sample sizes (n) are given for each insect performance parameter analyzed. Patterns indicate the following: white: chewers; black: suckers; cross-hatched: miners; left diagonal: gallers; right diagonal: borers.
P1: ARS
October 27, 1997
16:59
Annual Reviews
AR048-10
PLANT STRESS AND INSECT PERFORMANCE
201
P1: ARS
October 27, 1997
202
16:59
Annual Reviews
AR048-10
KORICHEVA ET AL
survival, and colonization were not affected. The fairly high fail-safe numbers (Table 1) indicate that the magnitude of the observed effects is robust enough to withstand the discovery of additional studies showing no effect. The survival of gall-makers was drastically reduced on stressed plants, as was their density of colonization. Although this conclusion is based on few studies, the result is robust, as indicated by the large fail-safe number of studies (Table 1). Miner performance was not significantly affected by plant stress, although fecundity, survival, and colonization densities tended to be higher on stressed plants. Finally, colonization densities of borers increased significantly on stressed trees. TIMING OF FEEDING White (123, 124) argued that the physiological responses of plants subjected to stress resemble changes associated with senescence, and, therefore, insects adapted to exploit senescing tissue (senescence feeders) will benefit from plant stress, whereas flush feeders may suffer if the stress reduces or terminates plant growth (also cf 95). Stress responses of feeding guilds that include exclusively flush (gall-makers) or senescence (most borers) feeders support White’s predictions. However, for feeding guilds representing both flush and senescence feeders (chewing and sucking insects), the timing of feeding had no effect; for chewing insects, no effects were found for RGR (d+ = −0.05 vs −0.07 for flush and senescence feeders, respectively; P = 0.94) or fecundity (d+ = −0.54 vs −0.47; P = 0.81); and for sucking insects, no effects were found for fecundity (d+ = 0.53 vs 0.44; P = 0.87). DEVELOPMENTAL STAGE White (123) argued that the stress-induced increase in availability of nitrogen should be most important for very young insects. This is in accordance with studies showing that early instars are generally more sensitive to plant quality than late instars are (84, 104), and that high mortality often occurs during early instars (93, 109, 120). Accordingly, stress-induced changes in plant quality may increase the probability of survival at this critical stage of the life cycle and thus have a profound effect on population dynamics (122). Data for testing this hypothesis are scarce because most bioassays have been conducted on late instars. Our analysis did not detect any significant differences (P = 0.55) among the mid instars (d+ = 0.003), late instars (d+ = −0.29), and the whole larval stage (d+ = −0.16) of chewing insects in terms of stressinduced effects on RGR. No comparable data were found for very young larvae. FEEDING SPECIALIZATION Jones & Coleman (52) suggested that insects specialized to feed on a few plant taxa should be limited in their physiological plasticity and, therefore, should be able to respond to altered plant nitrogen only over a narrow concentration range. In contrast, generalists may show greater behavioral plasticity and may therefore be able to handle a broad range
P1: ARS
October 27, 1997
16:59
Annual Reviews
AR048-10
PLANT STRESS AND INSECT PERFORMANCE
203
of nitrogen and secondary metabolite concentrations (52). However, our results indicate that the degree of specialization does not have any effect on the sign or the magnitude of responses of chewing insects to plant stress, as measured by RGR (d+ = −0.12 vs −0.08 for monophagous and polyphagous insects, respectively; P = 0.88) and fecundity (d+ = −0.50 vs −0.68, P = 0.54).
Variation Related to the Host Plant TYPE OF STRESS A basic premise of the PSH is that different types of abiotic stresses induce similar responses in plants, i.e. increased soluble nitrogen concentrations, and therefore, different stresses should affect insects similarly (123). However, White (123) did not regard shading as a stress factor and argued that insects should prefer sun leaves that senesce earlier and have higher concentrations of nitrogen than shaded leaves. In contrast, several other hypotheses predict that herbivory should be higher on shaded foliage (6, 46). In accordance with White (123), the type of stress applied to plants did not significantly modify responses of chewing insects in terms of RGR (P = 0.41) or fecundity (P = 0.32). However, there was a tendency for the RGR of chewers to increase in response to shading (d+ = 0.34, n = 3) although the effect was not significant (95% CI −0.304 to 1.02). Pollution increased the reproductive potential of sucking insects (d+ = 0.64, n = 21), whereas water stress tended to decrease their population growth (d+ = −0.31, n = 5); the difference among stress types was marginally significant (P = 0.07). Differences in fecundity responses between sucking and chewing insects were significant only in pollution experiments (d+ = 0.60 vs d+ = −0.68, P < 0.001), whereas water stress tended to decrease the fecundity of both types of insects (d+ = −0.27 vs d+ = −0.41, P = 0.73). Pollutant type had no significant effect on RGR or fecundity of sucking or chewing insects, although SO2 and heavy metals tended to have a more pronounced effect on insect performance as compared with ozone and acid rain effects. An increased reproductive potential of sucking insects in response to pollution was observed mainly in studies where exposure was direct (d+ = 0.95, n = 14). No such effects were detected when insects were introduced onto prefumigated plants or shielded from the exposure to pollutants in other ways (d+ = −0.04, n = 7). In contrast, the accumulation of pollutants in the host plant as well as direct exposure to pollutants have been shown to decrease the performance of chewing insects (40, 43, 44, 60). This finding indicates that the enhanced performance of sucking insects and the decreased performance of chewers may be due to the direct effects of pollutants on the insects rather than to plant-mediated effects. No significant difference existed between the effects of shading and of water stress on the colonization of plants by mining insects (P = 0.50). Responses
P1: ARS
October 27, 1997
204
16:59
Annual Reviews
AR048-10
KORICHEVA ET AL
of gall-makers and borers to different types of stresses could not be compared because they were studied only in water-stress experiments. PLANT GROWTH RATE Inherent growth rate of plants has been suggested as a predictor of resistance to stresses and susceptibility to herbivores (11, 19, 30). According to the predictions, fast growers under stress should display plasticity in secondary metabolite production owing to the high priority placed on allocation to growth. In slow growers, allocation to secondary compounds has a high priority, and therefore, secondary metabolite production should be relatively unaffected by stress (52). The effect of plant growth rate on responses of chewing insects to plant stress was significant only for survival (P = 0.010). However, there was a tendency for the plant stress–induced reductions of RGR, fecundity, and survival to be greater on slow-growing plants than on fast growers (Figure 2). Thus, it appears that chewing insects feeding on slowly growing plants suffer from plant stress more than their counterparts on fast-growing plants. The small number of studies hampered the analysis of the importance of plant growth rate for other insect guilds.
Figure 2 Mean effect sizes (d+) of stress treatments on the performance of chewing insects on plants with different relative growth rates (RGRs). Numbers adjacent to symbols show the sample size; vertical bars indicate the 95% confidence interval (CI).
P1: ARS
October 27, 1997
16:59
Annual Reviews
AR048-10
PLANT STRESS AND INSECT PERFORMANCE
205
Leaf life span is known to correlate negatively with plant growth rate (99), i.e. evergreen plants tend to grow more slowly than deciduous ones. We did not find any significant differences between stressed evergreen and stressed deciduous trees in the magnitude of insect responses in terms of RGR (both chewing and sucking insects) or fecundity (chewing insects).
LEAF LONGEVITY
In a recent review, Waring & Cobb (114) reported that the proportion of positive insect responses to water stress was higher on conifers than on angiosperms. Similiarly, we found that reproductive potential of sucking insects increased on stressed conifers (d+ = 0.70), whereas no significant changes were detected on deciduous trees (d+ = −0.13, P = 0.064). In contrast, the colonization by woodborers was enhanced by water stress more on angiosperms than on conifers (d+ = 1.93 vs 0.35, P = 0.001). We found no significant differences in responses of chewing insects to stress on angiosperms and conifers. PLANT TAXA
PLANT ONTOGENETIC STAGE For obvious practical reasons, most stress experiments involving trees (76%) have been carried out with seedlings and saplings instead of mature trees. There are, however, several factors that complicate attempts to extrapolate results obtained from seedlings to larger trees. In general, mature trees differ from seedlings in that they have a lower relative growth rate, allocate more resources to support tissue and reproduction, have access to more water and nutrients as a result of their more developed root systems, and have a greater capacity to store resources in stems and roots (55). As a result, seedlings may differ from mature trees of the same species in their susceptibility to abiotic stresses and insect herbivory (54, 61). Our analysis showed that decreases in fecundity of chewing insects tended to be more pronounced in experiments with saplings and mature trees compared with corresponding experiments with seedlings, but the difference was not statistically significant (P = 0.21). The reproductive potential of sucking insects tended to increase on stressed seedlings (d+ = 0.56) but not on stressed mature trees (d+ = −0.02), whereas colonization rates of borers were more enhanced in experiments with mature trees than in trials with saplings (d+ = 1.77 vs 0.53, P = 0.03).
Variation Related to Experimental Design Variation in effect sizes (d ) across studies could be largely due to sampling error if individual studies included in the analysis had few replicates (51). Most experiments aimed at testing the PSH had a two-level nested design: Each treatment was allocated to several trees, and bioassays with several insects were conducted on each tree. The number of trees per treatment ranged from 2 to 40 (median = 10), and the number of insects per tree varied
NUMBER OF REPLICATES
P1: ARS
October 27, 1997
206
16:59
Annual Reviews
AR048-10
KORICHEVA ET AL
Figure 3 Absolute magnitude of effect size (d ) of individual studies in relation to the total sample size (number of insects per treatment) used in the control or experimental group (whichever was smaller). Each point represents a single comparison. rs = −0.496, n = 151, P < 0.001.
from 1 to 75 (median = 6). Total sample size (number of insects per treatment) varied from 2 to 416 (median = 13). A clear tendency was apparent for effect sizes (d ) in experiments with total sample sizes less than 20 to vary more than those with larger samples (Figure 3). The same phenomenon was observed by Gurevitch et al (32) for competition experiments and is known as a “funnel effect” (68): As N increases, variation due to sampling error decreases, and the values “funnel” down toward the true effect size. Moreover, although effect sizes were weighted by a correction term to remove the bias due to small sample size, the absolute values of effect sizes decreased significantly with an increase in the number of trees per treatment (Spearman’s rank correlation, rs = −0.32, n = 134, P = 0.0002) and total sample size (Figure 3). Therefore, extreme positive and negative values for d reported in some experiments may have been the result of sampling error: 14 out of 15 experiments with absolute values of d > 2 had total sample sizes less than 20. DURATION AND TIMING OF STRESS TREATMENTS Mopper & Whitham (85) suggested that many discrepancies between the results of stress experiments and field observations can be explained by the short duration of stress treatments and by the fact that insect bioassays are conducted simultaneously with the plant stress treatment, whereas in nature insect outbreaks often occur after several years of stressful conditions. To test this hypothesis, insect responses in stress experiments where insect bioassays were conducted simultaneously with stress treatments were compared with those where the plant had been released from stress prior to the bioassay. No differences in insect responses were detected for the RGR (P = 0.45) or fecundity (P = 0.20) of chewing insects. Contrary to Mopper & Whitham’s prediction, the reproductive potential of sucking insects was increased by simultaneous stress (d+ = 0.64), whereas releasing plants from stress prior to bioassays tended to decrease their fecundity (d+ = −0.41,
P1: ARS
October 27, 1997
16:59
Annual Reviews
AR048-10
PLANT STRESS AND INSECT PERFORMANCE
207
Figure 4 Effect size of plant stress on fecundity of chewing insects as affected by the duration of stress treatment. Each point represents a single comparison.
P = 0.11). However, Mopper & Whitham’s hypothesis seemed to be supported by data on insect survival: Plant stress resulted in decreased survival of chewers when applied simultaneously with insect feeding (d+ = −0.80) but resulted in increased survival when the treatments were terminated before the bioassays (d+ = 0.34, P = 0.08). The duration of stress treatments in our database ranged from 5 h to 7 years, with a median of 4 months. There was no relation between the duration of the stress treatments and effect sizes (d ). Variation in insect responses was higher in short-term experiments than in long-term ones (Figure 4). TYPE OF EXPERIMENT The assumption seems reasonable that field experiments offer more realistic conditions under which to test stress-induced effects on insect performance compared with laboratory experiments. However, in our analysis, experiment type did not significantly affect the RGR (P = 0.86) or fecundity (P = 0.61) of chewing insects or the reproductive potential of sucking insects (P = 0.67). However, the fecundity of sucking insects increased significantly in response to plant stress in the laboratory experiments (d+ = 0.75) but not in the field. This discrepancy might have been due to the higher temperatures, and thus faster development of sucking insects, in the laboratory or to the influence of natural enemies that sometimes act as confounding factors in field experiments.
PATTERNS IN INSECT RESPONSES TO STRESSED PLANTS Our analysis provided very little support for the PSH in its most general, original form: In other words, plant stress had no overall effect on insect performance
P1: ARS
October 27, 1997
208
16:59
Annual Reviews
AR048-10
KORICHEVA ET AL
as measured in terms of insect growth rate, fecundity, survival, or colonization. However, we found large amounts of variation in the magnitude and direction of insect responses among stress experiments. This pattern suggests that plant stress/insect performance interactions form a continuum, ranging from situations in which the insect is favored by stressed plants to those in which it is favored by vigorous plants (63, 95). The analysis showed that insect feeding guild is the most important determinant of insect response and corroborated Larsson’s (63) prediction concerning the magnitude and sign of insect feeding guild responses to plant stress (also cf 119). This finding differs from a recent vote-count review by Waring & Cobb (114), who found little variation in the responses of different feeding guilds to water stress. This discrepancy could be due to the fact that Waring & Cobb included woodborers, leaf miners, and leaf chewers in the same guild, whereas we analyzed the responses of these insect groups separately. Moreover, we found that differences among guilds of chewers and suckers were less apparent in the case of water stress than in the case of pollution; Waring & Cobb (114) did not include pollution in their analysis. Subdividing the studies on the basis of insect feeding guild did not eliminate all of the variation in effect sizes (d ) within each group. Although some guilds were fairly consistent in their responses to plant stress (e.g. gallers and borers), chewers, suckers, and miners showed increased as well as decreased performance on stressed plants in different studies. Such within-guild variation may be due to differences in a variety of factors related to insect life history (inherent growth rate, mobility, aggregation), feeding style (diet breadth, tissue specialization), phenology, and taxonomic group (52). Timing of feeding and the degree of host specialization were the only factors for which there were enough data to carry out analyses. Significant within-guild differences were not detected between either flush- and senescence-feeding chewers and suckers or between monophagous and polyphagous chewers. This outcome could have been due partly to problems in categorizing some insect groups. For example, many aphid species shift between flushing and senescing tissues depending on their availability throughout the season (22, 62, 125), and therefore they cannot be classified as strictly flush or senescence feeders. Similarly, many leaf miners feed on flush tissue as early instars but on senescing tissues in later instars (124). However, not enough studies were available to subdivide these categories more finely. Earlier attempts to discern patterns with regard to the effect of plant stress on insects have tended to ignore variation due to plant characters. It is important to realize that the plant stress/insect performance interaction is a system characteristic, the nature of which is determined by both plant and insect. Therefore, the plant component needs to be considered in more detail. In our analysis, we were able to include four potentially important plant characteristics, i.e.
P1: ARS
October 27, 1997
16:59
Annual Reviews
AR048-10
PLANT STRESS AND INSECT PERFORMANCE
209
inherent growth rate, leaf longevity, ontogenetic stage, and taxonomic group. Minor effects were detected for most of these factors, but no consistent pattern emerged. Perhaps the most interesting result was that chewing insects were, in general, more favored by stressed fast-growing plants than by stressed slow-growing plants. More emphasis on plant variation would probably be rewarding in future research aimed at disentangling within-guild differences in stress response. For example, it would be interesting to conduct experiments where fast- and slow-growing plants were stressed by the same factor and the stress effects assessed using the same insect species (e.g. 69). A major premise of the original formulation of the PSH is that different kinds of stresses affect insect performance similarly (100, 123). Our analysis implies that this view may be overly simplistic, but the evidence was guild specific. No differences were detected between types of stresses with respect to chewing insects; their responses were negative regardless of stress type, except for a tendency for shading to have positive effects. In contrast, the performance of sucking insects was improved on polluted plants, whereas it was reduced on water-stressed ones. Direct exposure to pollutants may act as a confounding factor (40, 43, 44, 60). Future studies need to control for this effect, in cases where hypotheses concern stress-mediated effects on food quality.
STRESS EXPERIMENTS: PRESENT STATUS AND FUTURE DIRECTIONS There is good reason to assume that stress-induced effects in plants, and their effects on insect performance, are nonlinear (28, 63, 98). However, only 26% of the studies included in our analysis used several levels of stress, thereby allowing detection of any nonlinear relations. Several of these studies showed that insect performance improved with stress level until reaching some threshold, above which performance declined (60, 80, 92). Recent studies on bark beetles and their associated blue-stain fungi have revealed that mild drought stress actually increases tree resistance, thus indicating a nonlinear relationship between stress and the performance of these organisms (14, also cf 98). Stress experiments commonly include only one type of stress treatment. Effects of interactions among different types of stresses on insect performance were analyzed in only 29% of the experiments in this study. These experiments (87, 126, 117), and data from field studies (e.g. 65, 75), suggest the importance of the combined effects of drought conditions, nutrient deficiency, and/or air pollution. Thus, the single-factor approach commonly applied may be simplistic and may not fully simulate abiotic conditions thought to trigger insect outbreaks. Variation due to plant genotype was not taken into account in 69% of the studies that we reviewed. The few studies that included several plant genotypes
P1: ARS
October 27, 1997
210
16:59
Annual Reviews
AR048-10
KORICHEVA ET AL
when assessing plant stress effects on insect performance demonstrated that genotype-by-environment interactions can be quite common and that the magnitude—and even the sign—of insect response to plant stress may depend on plant genotype (17, 47, 73, 110). Our analysis did not include fertilization experiments. Such experiments have sometimes been considered as tests of nutrient deficiency, the rationale for which is that plants growing in habitats with poor soils (“chronic stress”) can be experimentally manipulated to represent “controls” through fertilization treatments (114). There are, however, conceptual problems associated with this approach. If one is interested in testing the plant stress hypothesis (PSH) in order to predict propensity of insect outbreaks, processes need to be considered on an ecological time scale. However, organisms living in chronic-stress habitats are likely to have evolved adaptations to cope with adverse abiotic conditions (11, 19, 30). It would be interesting to explore the possibility that sensitivity to short-term stress differs between plants growing in chronic-stress situations and conspecifics in richer habitats. The results of stress experiments can be interpreted correctly only if the treatment effects can be properly assessed. Owing to the complex nature of the effects of stressful abiotic conditions on plants (67), natural stress is not easily simulated in experiments. In particular, the physiological status of plants designated as “stressed” in experimental situations or in the field is often not known. Insect outbreaks thought to be triggered by plant stress are often discovered too late to be able to characterize the putative stress in the host plant. In not a single one of the studies included in this review was the stress level related to the physiology or internal chemistry of the same plant species subjected to natural stress. Although this drawback is typical of studies in which abiotic factors are manipulated, it should, nevertheless, be considered a good reason to interpret the data with caution. The ultimate goal of experiments on plant stress/insect performance is to determine whether plant stress is likely to have an impact on the dynamics of insect herbivore populations. The large number of findings indicating correlations between stressed plants and insect outbreaks (e.g. 75) is not easily interpreted because stressful abiotic conditions can have direct effects on the herbivores (13), as well as indirect effects mediated by both host plants and natural enemies (36). The vagueness of experimental support for the PSH, documented in this and earlier reviews (63, 86, 114, 119), is remarkable considering the rather large number of experiments conducted. It is encouraging that some patterns are emerging, particularly in relation to insect feeding guild and, to a lesser extent, stress type. However, a great deal of variation remains to be explained. Some of this variation can probably be resolved by designing better replicated experiments, by studying different levels of stress (thereby allowing
P1: ARS
October 27, 1997
16:59
Annual Reviews
AR048-10
PLANT STRESS AND INSECT PERFORMANCE
211
detection of nonlinear relationships in responses of insects to stressed plants), or by including a variety of stressors in order to elucidate interaction effects. We believe, however, that the greatest emphasis should be placed on obtaining a detailed understanding of how plants respond to abiotic stress, particularly from the perspective of the insect under study. Many insect species feed specifically on resources that are variable in the plant, e.g. tissues with different functions, such as buds, leaves, phloem, and roots, or the same tissue types that vary according to their sink/source status (38, 76). These different feeding sites are likely to be differentially affected by stressful conditions. For instance, some tissues may be more sensitive than others, which may, in turn, translate into dissimilar insect responses. Thus, a better understanding of feeding site characteristics and their responses to abiotic stress would improve our ability to explain the variation in insect responses to stressed plants that is evident in the literature. A cooperative effort between entomologists and plant physiologists will therefore be required to reformulate the PSH successfully. The complexities of the plant stress/insect performance interactions can be understood only if knowledge about both components of the system is appreciated. ACKNOWLEDGMENTS S H¨oglund, E Kula, T Virtanen, and J Vranjic helped with our literature search and data collection. Ø Austar˚a, C Bj¨orkman, DA Herms, ID Hodkinson, J Holopainen, PC Marino, S Neuvonen, DA Potter, RW Preszler, PW Price, DW Ross, K Ruohom¨aki, LM Schroeder, AD Watt, and JB Whittaker generously provided additional data not available in their articles. C Bj¨orkman, RF Denno, B Ekbom, C Glynn, J Gurevitch, T Honkanen, A Huberty, MV Kozlov, WJ Mattson, D Quiring, and D Tilles commented on early versions of the manuscript. MS Rosenberg shared his expert help on using the MetaWin statistical package. We are grateful to all of these people. The study was financially supported by grants from the Swedish Natural Science Research Council, the Swedish Council of Forestry and Agricultural Research, and the Academy of Finland. Visit the Annual Reviews home page at http://www.AnnualReviews.org.
Literature Cited 1. Adams DC, Gurevitch J, Rosenberg MS. 1997. Resampling tests for metaanalysis of ecological data. Ecology. 78:1277–83 2. Arnqvist G, Wooster D. 1995. Metaanalysis: synthesizing research findings
in ecology and evolution. TREE 10:236– 40 3. Austar˚a Ø, Christiansen E. 1996. Drought stress and Neodiprion sertifer dynamics. Res. Pap. Nor. For. Res. Inst. 16 pp. (In Norwegian)
P1: ARS
October 27, 1997
212
16:59
Annual Reviews
AR048-10
KORICHEVA ET AL
4. Austar˚a Ø, Midtgaard F. 1987. A preliminary experiment on the fecundity of Neodiprion sertifer on Pinus sylvestris grown in forest soil of various acidity. Scand. J. For. Res. 2:365–67 5. Braun S, Fl¨uckiger W. 1989. Effect of ambient ozone and acid mist on aphid development. Environ. Pollut. 56:177– 87 6. Bryant JP, Chapin FS III, Klein DR. 1983. Carbon/nutrient balance of boreal plants in relation to vertebrate herbivory. Oikos 40:357–68 7. Bultman TL, Faeth SH. 1987. Impact of irrigation and experimental drought stress on leaf-mining insects of Emory oak. Oikos 48:5–10 8. Bultman TL, Faeth SH. 1988. Abundance and mortality of leaf miners on artificially shaded Emory oak. Ecol. Entomol. 13:131–42 9. Cannon WN Jr. 1993. Gypsy moth (Lepidoptera: Lymantriidae) consumption and utilization of northern red oak and white oak foliage exposed to simulated acid rain and ozone. Environ. Entomol. 22:669–73 10. Carroll AL, Quiring DT. 1993. Interactions between size and temperature influence fecundity and longevity of a tortricid moth, Zeiraphera canadensis. Oecologia 93:233–41 11. Chapin FS III, Autumn K, Pugnaire F. 1993. Evolution of suites of traits in response to environmental stress. Am. Nat. 142:S78–92 12. Christiansen E, Austar˚a Ø. 1996. Development of Neodiprion sertifer (Geoff) (Hym., Diprionidae) on droughtstressed pines: a laboratory experiment. J. Appl. Entomol. 120:221–23 13. Christiansen E, Bakke A. 1997. Does drought really enhance Ips typographus epidemics?—A Scandinavian perspective. In Proceedings: Integrating Cultural Tactics into the Management of Bark Beetle and Reforestation Pests, ed. JC Gregoire, AM Liebhold, FM Stephen, KR Day, SM Salom. USDA For. Serv. Gen. Tech. Rep. In press 14. Christiansen E, Glosli AM. 1996. Mild drought enhanced the resistance of Norway spruce to a bark beetle-transmitted fungus. See Ref. 76a, pp. 192–99 15. Cockfield SD, Potter DA. 1986. Interaction of euonymus scale (Homoptera: Diaspididae) feeding damage and severe water stress on leaf abscission and growth of Euonymus fortunei. Oecologia 71:41–46 16. Cohen J. 1988. Statistical Power Anal-
17.
18.
19. 20.
21.
22.
23.
24. 25. 26.
27.
28.
29.
ysis for the Behavioral Sciences. Hillsdale: Erlbaum. 567 pp. Coleman JS, Jones CG. 1988. Acute ozone stress on eastern cottonwood (Populus deltoides Bartr.) and the pest potential of the aphid, Chaitophorus populicola Thomas (Homoptera: Aphididae). Environ. Entomol. 17:207–12 Coleman JS, Jones CG. 1988. Plant stress and insect performance: cottonwood, ozone and a leaf beetle. Oecologia 76:57–61 Coley PD, Bryant JP, Chapin FS III. 1985. Resource availability and plant antiherbivore defense. Science 230:895–99 Craig TP, Price PW, Itami JK. 1992. Facultative sex ratio shifts by a herbivorous insect in response to variation in host plant quality. Oecologia 92:153–61 Csada RD, James PC, Espie RHM. 1996. The “file drawer problem” of nonsignificant results: Does it apply to biological research? Oikos 76:591–93 Dixon AFG. 1970. Quality and availability of food for a sycamore aphid population. In Animal Populations in Relation to Their Food Resources, ed. A Watson, pp. 271–87. Oxford: Blackwell Dixon AFG, Wratten SD. 1971. Laboratory studies on aggregation, size and fecundity in the black bean aphid, Aphis fabae Scop. Bull. Entomol. Res. 61:97– 111 Dohmen GP. 1985. Secondary effects of air pollution: enhanced aphid growth. Environ. Pollut. 39:227–34 Dudt JF, Shure DJ. 1994. The influence of light and nutrients on foliar phenolics and insect herbivory. Ecology 75:86–98 Dunn JP, Kimmerer TW, Nordin GL. 1986. The role of host tree condition in attack of white oaks by the twolined chestnut borer, Agrilus bilineatus (Weber) (Coleoptera: Buprestidae). Oecologia 70:596–600 Dunn JP, Lorio PL. 1993. Modified water regimes affect photosynthesis, xylem water potential, cambial growth, and resistance of juvenile Pinus taeda L. to Dendroctonus frontalis (Coleoptera: Scolytidae). Environ. Entomol. 22:948– 57 English-Loeb GM. 1989. Nonlinear responses of spider mites to droughtstressed host plants. Ecol. Entomol. 14:45–55 Foggo A, Speight MR. 1993. Root damage and water stress: treatments affecting the exploitation of the buds of common ash Fraxinus excelsior L., by larvae of the ash bud moth Prays fraxinella
P1: ARS
October 27, 1997
16:59
Annual Reviews
AR048-10
PLANT STRESS AND INSECT PERFORMANCE
30.
31.
32.
33.
34.
35.
36.
37.
38. 39.
40.
41.
42.
Bjerk. (Lep., Yponomeutidae). Oecologia 96:134–38 Grime JP. 1977. Evidence for the existence of three primary strategies in plants and its relevance to ecological and evolutionary theory. Am. Nat. 111:1169– 94 Gurevitch J, Hedges LV. 1993. Metaanalysis: combining the results of independent experiments. In Design and Analysis of Ecological Experiments, ed. SM Scheiner, J Gurevitch, pp. 378–425. New York: Chapman & Hall Gurevitch J, Morrow LL, Wallace A, Walsh JS. 1992. A meta-analysis of competition in field experiments. Am. Nat. 140:539–72 Haack RA, Benjamin DM. 1982. The biology and ecology of the twolined chestnut borer, Agrilus bilineatus (Coleoptera: Buprestidae), on oaks, Quercus spp., in Wisconsin. Can. Entomol. 114:385–96 Hanks LM, Denno RF. 1993. Natural enemies and plant water relations influence the distribution of an armored scale insect. Ecology 74:1081–91 Hanks LM, Paine TD, Millar JG. 1991. Mechanisms of resistance in Eucalyptus against larvae of the eucalyptus longhorned borer (Coleoptera: Cerambycidae). Environ. Entomol. 20:1583– 88 Hanski I. 1990. Small mammal predation and the population dynamics of Neodiprion sertifer. See Ref. 119a, pp. 253–64 Hartley SE, Gardner SM. 1995. The response of Philaenus spumarius (Homoptera: Cercopidae) to fertilizing and shading its moorland host-plant (Calluna vulgaris). Ecol. Entomol. 20:396– 99 Haukioja E, Honkanen T. 1996. Why are tree responses to herbivory so variable? See Ref. 76a, pp. 1–10 Haukioja E, Neuvonen S. 1985. The relationship between size and reproductive potential in male and female Epirrita autumnata (Lep., Geometridae). Ecol. Entomol. 10:267–70 Heli¨ovaara K, V¨ais¨anen R. 1989. Reduced cocoon size of diprionids (Hymenoptera) reared on pollutant affected pines. J. Appl. Entomol. 107:32–40 Heli¨ovaara K, V¨ais¨anen R. 1990. Prolonged development in Diprion pini (Hymenoptera, Diprionidae) reared on pollutant affected pines. Scand. J. For. Res. 5:127–31 Heli¨ovaara K, V¨ais¨anen R. 1993. Insects
43.
44.
45.
46. 47.
48.
49.
50.
51.
52.
213
and Pollution. Boca Raton: CRC Press. 393 pp. Heli¨ovaara K, V¨ais¨anen R, Kemppi E. 1989. Change of pupal size of Panolis flammea (Lepidoptera: Noctuidae) and Bupalus piniarius (Geometridae) in response to concentration of industrial pollutants in their food plant. Oecologia 79:179–83 Heli¨ovaara K, V¨ais¨anen R, Varama M. 1989. Cocoon measurements of gregarious pine sawflies (Hymenoptera, Diprionidae) reared on pollutant-stressed Scots pines. Ann. Entomol. Fenn. 55:75– 78 Heli¨ovaara K, V¨ais¨anen R, Varama M. 1990. Fecundity and egg viability in relation to female body size in Neodiprion sertifer (Hymenoptera: Diprionidae). Holarct. Ecol. 13:166–72 Herms DA, Mattson WJ. 1992. The dilemma of plants: to grow or to defend. Q. Rev. Biol. 67:238–335 Herms DA, Mattson WJ, Karowe DN, Coleman MD, Trier TM, et al. 1996. Variable performance of outbreak defoliators on aspen clones exposed to elevated CO2 and O3. In Proc. 1995 Meet. North. Glob. Change Prog., Pittsburgh, PA, ed. J Hom, R Birdsey, K O’Brian, pp. 43–55. USDA For. Serv. NE For. Exp. Stn. Gen. Tech. Rep. NE-214 Holopainen J, Braun S, Fl¨uckiger W. 1994. The response of spruce shoot aphid Cinara pilicornis Hartig to ambient and filtered air at two elevations and pollution climates. Environ. Pollut. 86:233–38 Holopainen JK, Kainulainen E, Oksanen J, Wulff A, K¨arenlampi L. 1991. Effect of exposure to fluoride, nitrogen compounds and SO2 on the numbers of spruce shoot aphids on Norway spruce seedlings. Oecologia 86:51–56 Holopainen JK, Mustaniemi A, Kainulainen P, Satka H, Oksanen J. 1993. Conifer aphids in an air-polluted environment. I. Aphid density, growth and accumulation of sulphur and nitrogen by Scots pine and Norway spruce seedlings. Environ. Pollut. 80:185–91 Hunter JE, Schmidt FL. 1990. Methods of Meta-Analysis: Correcting Error and Bias in Research Findings. Newbury Park: Sage. 592 pp. Jones CG, Coleman JS. 1991. Plant stress and insect herbivory: toward an integrated perspective. In Response of Plants to Multiple Stresses, ed. HA Mooney, WE Winner, EJ Pell, pp. 249– 80. San Diego: Academic
P1: ARS
October 27, 1997
214
16:59
Annual Reviews
AR048-10
KORICHEVA ET AL
53. Kainulainen P, Holopainen JK, Hyttinen H, Oksanen J. 1994. Effect of ozone on the biochemistry and aphid infestation of Scots pine. Phytochemistry 35:39– 42 54. Kearsley MJC, Whitham TG. 1989. Developmental changes in resistance to herbivory: implications for individuals and populations. Ecology 70:422–34 55. Kelly JM, Samuelson L, Edwards G, Hanson P, Kelting D, et al. 1995. Are seedlings reasonable surrogates for trees? An analysis of ozone impact on Quercus rubra. Water Air Soil Pollut. 85:1317–24 56. Kennedy JS, Lamb KP, Booth CO. 1958. Responses of Aphis fabae Scop. to water shortage in host plants in pots. Entomol. Exp. Appl. 1:274–91 57. Kidd NAC. 1990. The effects of simulated acid mist on the growth rates of conifer aphids and the implications for tree health. J. Appl. Entomol. 110:524– 29 58. Kile GA, Elliot HJ, Candy SG, Hall MF. 1992. Treatments influencing susceptibility of Nothofagus cunninghamii to the ambrosia beetle Platypus subgranosus in Australia. Can. J. For. Res. 22:769– 75 59. Kimberling DN, Scott ER, Price PW. 1990. Testing a new hypothesis: plant vigor and phyloxera distribution on wild grape in Arizona. Oecologia 84:1–8 60. Kozlov MV, Zvereva EL, Selikhovkin AV. 1996. Decreased performance of Melasoma lapponica (Coleoptera: Chrysomelidae) fumigated by sulphur dioxide: direct toxicity versus host plant quality. Environ. Entomol. 25:143–46 61. Kozlowski TT, Kramer PJ, Pallardy SG. 1991. The Physiological Ecology of Woody Plants. San Diego: Academic 62. Larsson S. 1985. Seasonal changes in the within-crown distribution of the aphid Cinara pini on Scots pine. Oikos 45:217–22 63. Larsson S. 1989. Stressful times for the plant stress-insect performance hypothesis. Oikos 56:277–83 64. Larsson S, Bj¨orkman C. 1993. Performance of chewing and phloem-feeding insects on stressed trees. Scand. J. For. Res. 8:550–59 65. Larsson S, Tenow O. 1984. Areal distribution of Neodiprion sertifer (Hym., Diprionidae) outbreak on Scots pine as related to stand condition. Holarct. Ecol. 7:81–90 66. Leather SR. 1988. Size, reproductive potential and fecundity in insects: Things
67. 68. 69.
70. 71.
72.
73.
74.
75.
76.
76a.
77.
aren’t as simple as they seem. Oikos 51:386–88 Levitt J. 1980. Responses of Plants to Environmental Stresses, Vol. 1. New York: Academic Light RJ, Pillemer DB. 1984. Summing Up: the Science of Reviewing Research. Cambridge: Harvard Univ. Press Lindroth RL, Reich PB, Tjoelker MG, Volin JC, Oleksyn J. 1993. Light environment alters response to ozone stress in seedlings of Acer saccharum Marsh. and hybrid Populus L. III. Consequences for performance of gypsy moth. New Phytol. 124:647–51 Loehle C. 1988. Tree life history strategies: the role of defenses. Can. J. For. Res. 18:209–22 Lorio PL, Hodges JD. 1977. Tree water status affects induced southern pine beetle attack and brood production. USDA For. Res. Pap. SO-135. South. For. Exp. Stn. Major EJ. 1990. Water stress in Sitka spruce and its effect on the green spruce aphid Elatobium abietinum. See Ref. 119a, pp. 85–93 Marino PC, Cornell HV, Kahn DH. 1993. Environmental and clonal influences on host choice and larval survival in a leafmining insect. J. Anim. Ecol. 62:503–10 Martinat PJ. 1987. The role of climatic variation and weather in forest insect outbreaks. In Insect Outbreaks, ed. P Barbosa, JC Schultz, pp. 241–68. San Diego: Academic Mattson WJ, Haack RA. 1987. The role of drought stress in provoking outbreaks of phytophagous insects. In Insect Outbreaks, ed. P Barbosa, JC Schultz, pp. 365–407. San Diego: Academic Mattson WJ, Lawrence RK, Haack RA, Herms DA, Charles PJ. 1988. Plant defensive strategies for different insect feeding guilds in relation to plant ecological strategies and intimacy of host association. In Mechanisms of Plant Resistance to Insects, ed. WJ Mattson, J Levieux, C Bernard-Dagan, pp. 1–38. New York: Springer-Verlag Mattson WJ, Niemel¨a P, Rousi M. 1996. Dynamics of Forest Herbivory: Quest for Pattern and Principle. USDA For. Serv. Gen. Tech. Rep. NC-183 Mattson WJ, Slocum SS, Koller CN. 1983. Spruce budworm (Choristoneura fumiferana) performance in relation to foliar chemistry of its host plants. In Proceedings Forest Defoliator–Host Interactions: A Comparison Between Gypsy
P1: ARS
October 27, 1997
16:59
Annual Reviews
AR048-10
PLANT STRESS AND INSECT PERFORMANCE
78.
79.
80. 81.
82.
83.
84.
85.
86.
87.
88.
89.
90.
Moth and Spruce Budworm, ed. RL Talerico, M Montgomery, pp. 55–65. USDA For. Serv. Gen. Tech. Rep. NE85 McCullough DG, Wagner MR. 1987. Influence of watering and trenching ponderosa pine on a pine sawfly. Oecologia 71:382–87 McMillin JD, Wagner MR. 1995. Season and intensity of water stress: hostplant effects on larval survival and fecundity of Neodiprion gillettei (Hymenoptera: Diprionidae). Environ. Entomol. 24:1251–57 McNeill S, Whittaker JB. 1990. Air pollution and tree-dwelling aphids. See Ref. 119a, pp. 195–208 Miles PW, Aspinall D, Correll AT. 1982. The performance of two chewing insects on water-stressed food plants in relation to changes in their chemical composition. Aust. J. Zool. 30:347–55 Miller RH, Whitney HS, Berryman AA. 1986. Effects of induced translocation stress and bark beetle attack (Dendroctonus ponderosae) on heat pulse velocity and the dynamic wound response of lodgepole pine (Pinus contorta var. latifolia). Can. J. Bot. 64:2669–74 Mitterb¨ock F, F¨uhrer E. 1988. Wirkungen fluorbelasteter Fichtennadeln auf Nonnenraupen, Lymantria monacha L. (Lep., Lymantriidae). J. Appl. Entomol. 105:19–27 Montgomery ME. 1982. Life-cycle nitrogen budget for the gypsy moth, Lymantria dispar, reared on artificial diet. J. Insect Physiol. 28:437–42 Mopper S, Whitham TG. 1992. The plant stress paradox: effects on pinyon sawfly sex ratios and fecundity. Ecology 73:515–25 Myers JH. 1988. Can a general hypothesis explain population cycles of forest Lepidoptera? Adv. Ecol. Res. 18:179– 242 Neuvonen S, Lindgren M. 1987. The effect of simulated acid rain on performance of the aphid Euceraphis betulae (Koch) on silver birch. Oecologia 74:77–80 Neuvonen S, Lindgren M, Suomela J. 1987. The effects of artificial acid rain on the quality of birch leaves for aphids. Aquilo Ser. Bot. 25:109–14 (In Finnish) Neuvonen S, Routio I, Haukioja E. 1992. Combined effects of simulated acid rain and aphid infestation on the growth of Scots pine (Pinus sylvestris) seedlings. Ann. Bot. Fenn. 29:101–6 Neuvonen S, Saikkonen K, Suomela J.
91.
92.
93.
94.
95. 96.
97.
98.
99.
100.
101.
215
1990. Effect of simulated acid rain on the growth performance of the European pine sawfly (Neodiprion sertifer). Scand. J. For. Res. 5:541–50 Neuvonen S, Virtanen T, Ranta H, Nikula A. 1992. The pest and disease resistance of Scots pine and sulphur and heavy metal deposition. In The Lapland Forest Damage Project. Interim Report, ed. H Kauhanen, M Varmola, pp. 186– 93. Rovaniemi: Finn. For. Res. Inst. Res. Pap. 413 (In Finnish) Potter DA. 1992. Abundance and mortality of a specialist leafminer in response to experimental shading and fertilization of American holly. Oecologia 91:14–22 Preszler RW, Price PW. 1988. Host quality and sawfly populations: a new approach to life table analysis. Ecology 69:2012–20 Preszler RW, Price PW. 1995. A test of plant-vigor, plant-stress, and plantgenotype effects on leaf-miner oviposition and performance. Oikos 74:485– 92 Price PW. 1991. The plant vigor hypothesis and herbivore attack. Oikos 62:244– 51 Price PW, Clancy KM. 1986. Multiple effects of precipitation on Salix lasiolepis and populations of the stemgalling sawfly, Euura lasiolepis. Ecol. Res. 1:1–14 Redak RA, Paine TD, Trumble JT. 1995. Trirhabda geminata (Coleoptera: Chrysomelidae) resistance to the direct impact of simulated acidic fog on larval growth and mortality. Environ. Pollut. 90:61–66 Reeve JD, Ayres MP, Lorio PL. 1995. Host suitability, predation, and bark beetle population dynamics. In Population Dynamics: New Approaches and Synthesis, ed. N Cappuccino, PW Price, pp. 339–57. San Diego: Academic Reich PB, Walters MB, Ellsworth DS. 1992. Leaf life-span in relation to leaf, plant, and stand characteristics among diverse ecosystems. Ecol. Monogr. 62:365–92 Rhoades DF. 1979. Evolution of plant chemical defense against herbivores. In Herbivores: Their Interaction with Secondary Plant Metabolites, ed. GA Rosenthal, DH Janzen, pp. 3–54. New York: Academic Rosenberg MS, Adams DC, Gurevitch J. 1997. MetaWin: Statistical Software for Meta-Analysis with Resampling Tests. Sunderland: Sinauer
P1: ARS
October 27, 1997
216
16:59
Annual Reviews
AR048-10
KORICHEVA ET AL
102. Rosenthal R. 1979. The “file drawer problem” and tolerance for null results. Psychol. Rep. 12:491–511 103. Ross DW, Berisford CW. 1990. Nantucket pine tip moth (Lepidoptera: Tortricidae) response to water and nutrient status of loblolly pine. For. Sci. 36:719– 33 104. Rossiter M, Gershenzon J, Marby TJ. 1986. Behavioral and growth responses of specialist herbivore Homoeosoma electellum to major terpenoid of its host, Helianthus spp. J. Chem. Ecol. 12:1505– 21 105. Ruohom¨aki K, Chapin FS III, Haukioja E, Neuvonen S, Suomela J. 1996. Delayed inducible resistance in mountain birch in response to fertilization and shade. Ecology 77:2302–11 106. Saikkonen K, Neuvonen S, Kainulainen P. 1995. Oviposition and larval performance of European pine sawfly in relation to irrigation, simulated acid rain and resin acid concentration in Scots pine. Oikos 74:273–82 107. Salt DT, Whittaker JB. 1995. Populations of root-feeding aphids in the Liphook forest fumigation experiment. Plant Cell Environ. 18:321–25 108. Sj¨odin K, Schroeder LM, Eidmann HH, Norin T, Wold S. 1989. Attack rates of scolytids and composition of volatile wood constituents in healthy and mechanically weakened pine trees. Scand. J. For. Res. 4:379–91 109. Sone K. 1987. Population dynamics of the pine needle gall midge, Thecodiplosis japonensis Uchida et Inouye (Diptera, Cecidomyiidae). J. Appl. Entomol. 103:386–402 110. Talhouk SN, Nielsen DG, Montgomery ME. 1990. Water deficit, defoliation, and birch clones: short-term effect on gypsy moth (Lepidoptera: Lymantriidae) performance. Environ. Entomol. 19:937–42 111. Thomas AT, Hodkinson ID. 1991. Nitrogen, water stress and the feeding efficiency of lepidopteran herbivores. J. Anim. Ecol. 28:703–20 112. Trumbule RB, Denno RF. 1995. Light intensity, host-plant irrigation, and habitat-related mortality as determinants of the abundance of azalea lace bug (Heteroptera: Tingidae). Environ. Entomol. 24:898–908 113. Wagner MR, Frantz DP. 1990. Influence of induced water stress in ponderosa pine on pine sawflies. Oecologia 83:452–57 114. Waring GL, Cobb NS. 1992. The impact of plant stress on herbivore population
115.
116.
117.
118.
119.
119a.
120.
121.
122.
123.
124.
125.
126.
dynamics. In Insect-Plant Interactions, ed. EA Bernays, 4:167–226. Boca Raton: CRC Press Waring GL, Price PW. 1988. Consequences of host plant chemical and physical variability to an associated herbivore. Ecol. Res. 3:205–16 Waring RH, Pitman GB. 1985. Modifying lodgepole pine stands to change susceptibility to mountain pine beetle attack. Ecology 66:889–97 Warrington S, Whittaker JB. 1990. Interactions between Sitka spruce, the green spruce aphid, sulphur dioxide pollution and drought. Environ. Pollut. 65:363–70 Watt AD. 1986. The performance of the pine beauty moth on water-stressed lodgepole pine plants: a laboratory experiment. Oecologia 70:578–79 Watt AD. 1994. The relevance of the stress hypothesis to insects feeding on tree foliage. In Individuals, Populations and Patterns in Ecology, ed. SR Leather, AD Watt, NJ Mills, KFA Walters, pp. 73–85. Andover, UK: Intercept Watt AD, Leather SR, Hunter MD, Kidd NAC, eds. 1990. Population Dynamics of Forest Insects. Andover, UK: Intercept West C. 1985. Factors underlying the late seasonal appearance of the lepidopterous leaf-mining guild on oak. Ecol. Entomol. 10:111–20 White TCR. 1969. An index to measure weather-induced stress of trees associated with outbreaks of psyllids in Australia. Ecology 50:905–9 White TCR. 1974. A hypothesis to explain outbreaks of looper caterpillars, with special reference to populations of Selidosema suavis in a plantation of Pinus radiata in New Zealand. Oecologia 16:279–301 White TCR. 1984. The abundance of invertebrate herbivores in relation to the availability of nitrogen in stressed food plants. Oecologia 63:90–105 White TCR. 1993. The Inadequate Environment: Nitrogen and the Abundance of Animals. Berlin: Springer-Verlag. 425 pp. Wratten SD. 1974. Aggregation in the birch aphid Euceraphis punctipennis (Zett) in relation to food quality. J. Anim. Ecol. 43:191–98 Young CE, Hall RW. 1986. Factors influencing suitability of elms for elm leaf beetle, Xanthogaleruca luteola (Coleoptera: Chrysomelidae). Environ. Entomol. 15:843–49
