SEARCH

SEARCH BY CITATION

Keywords:

  • Chrysomelidae;
  • host plant;
  • Leptinotarsa;
  • life history correlations;
  • niche width;
  • plant–insect interaction;
  • Solanum;
  • specialization

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References
  9. Supporting Information

The genetic basis of host plant use by phytophagous insects can provide insight into the evolution of ecological niches, especially phenomena such as specialization and phylogenetic conservatism. We carried out a quantitative genetic analysis of multiple host use traits, estimated on five species of host plants, in the Colorado potato beetle, Leptinotarsa decemlineata (Coleoptera: Chrysomelidae). Mean values of all characters varied among host plants, providing evidence that adaptation to plants may require evolution of both behavioral (preference) and post-ingestive physiological (performance) characteristics. Significant additive genetic variation was detected for several characters on several hosts, but not in the capacity to use the two major hosts, a pattern that might be caused by directional selection. No negative genetic correlations across hosts were detected for any ‘performance’ traits, i.e. we found no evidence of trade-offs in fitness on different plants. Larval consumption was positively genetically correlated across host plants, suggesting that diet generalization might evolve as a distinct trait, rather than by independent evolution of feeding responses to each plant species, but several other traits did not show this pattern. We explored genetic correlations among traits expressed on a given plant species, in a first effort to shed light on the number of independent traits that may evolve in response to selection for host–plant utilization. Most traits were not correlated with each other, implying that adaptation to a novel potential host could be a complex, multidimensional ‘character’ that might constrain adaptation and contribute to the pronounced ecological specialization and the phylogenetic niche conservatism that characterize many clades of phytophagous insects.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References
  9. Supporting Information

The history of adaptive radiations, and thus of much of the diversification of life, is largely a history of shifts to new ecological niches: the utilization of new habitats or resources (Schluter, 2000). Phytophagous insects offer abundant opportunities to address questions about the evolution of niches. Many of the 400 000 described species are specialists, associated with a single family or even a single species of host plant (Bernays & Chapman, 1994; Schoonhoven et al., 2005). Phylogenetic conservatism is common: related insects often (although by no means always) feed on confamilial or even more closely related plants (Brues, 1924; Ehrlich & Raven, 1964; Mitter et al., 1991). The leaf beetle genus Leptinotarsa illustrates these patterns. Almost all the species with known hosts have been recorded from only one or two species of plants (Jacques, 1988), with the exception of the relative generalist Leptinotarsa decemlineata (Colorado potato beetle), which has been recorded from at least 20 plant species, most of which have been adopted only within the last 2 centuries. Host association is quite phylogenetically conservative, in that most species feed on Solanaceae, although other Leptinotarsa species are known to specialize on plants in the Asteraceae or Zygophyllaceae. (The phylogeny of Leptinotarsa has not yet been determined.)

These patterns pose two salient questions: Why is the dietary niche often so specialized (Futuyma & Moreno, 1988)? And what accounts for phylogenetic niche conservatism (Thompson, 1996; Wiens & Graham, 2005)? Both questions reflect the more fundamental problem of what restricts an insect species from incorporating new plant species into its diet, which could result in either niche expansion (polyphagy) or a shift to novel plant taxa. Although it has commonly been supposed that the rate and direction of evolution is seldom limited by availability of genetic variation, recent opinion appears to be shifting toward the hypothesis that genetic biases or constraints may be important in limiting responses to selection, especially on complex life history or ecological traits such as food or habitat use (Bradshaw, 1991; Futuyma et al., 1995; Schluter, 1996; Blows & Hoffmann, 2005). Such genetic limitations may be manifested in several related ways. First, some characters display little or no detectable genetic variation (Bradshaw, 1991; Hoffmann et al., 2003). Second, negative genetic correlations between pairs of traits are a widely appreciated potential constraint on the evolution of life history traits and other characters (Via & Lande, 1985; Barton & Turelli, 1989). Third, the response to selection may be lower, the greater the number of genetically independent components an adaptation comprises (Fisher, 1930; Wagner, 1988; Orr, 2000). The response to selection may be slight even if all component traits are genetically variable (Blows et al., 2004), because there may exist little or no genetic variation along the multivariate direction of selection (Schluter, 1996; Blows & Hoffmann, 2005). In contrast, genetic correlations among component traits, by reducing the dimensionality of genetic variation, may increase ‘evolvability’. For example, response to selection on body size would be greater if the growth of all body parts were coordinated than if many organs were genetically independent and had to evolve individually (Wagner & Altenberg, 1996; Orr, 2000).

The evolution of host associations in phytophagous insects may be affected by all these factors. Quantitative genetic screening of four species of Ophraella leaf beetles revealed genetic variation in responses to only some of the novel potential host plants with which they were challenged (Futuyma et al., 1995; Keese, 1998). Negative correlations in performance across plant species, which might favour host specialization (Dethier, 1954; Futuyma & Moreno, 1988; Jaenike, 1990), have been reported in a few studies (Fry, 1990; MacKenzie, 1996; Tilmon et al., 1998), although nonsignificant or even positive genetic correlations have been reported far more often (Fry, 1996; Berlocher & Feder, 2002; Futuyma, in press).

The dimensionality of adaptation to a host plant – the genetic independence or correlation among the components of adaptation – has been less studied, although it often comprises life history features such as phenology, behaviour (‘preference’ for or ‘acceptance’ of host plants), and physiological and/or morphological adaptations that enhance growth and survival (common measures of ‘performance’). However, many studies of ‘performance’, measured by indices such as larval survival or weight gain, do not distinguish the contributions of feeding rate (which may reflect ‘preference’) from post-ingestive physiological effects. Thus, the significance of the few cases of genetic correlation between ‘preference’ and ‘performance’ that have been described (e.g. Ng, 1988; Singer et al., 1988; Bossart, 2003) may be uncertain. A few studies have indicated that preference and performance may be genetically independent (e.g. Thompson, 1988; Thompson & Pellmyr, 1991; Wade, 1994; Forister, 2005). Moreover, both ‘preference’ and ‘performance’ are themselves complex characters (Feeny, 1992; Bernays & Chapman, 1994), and even the degree to which larval feeding, adult feeding, and oviposition responses are genetically different characters in holometabolous insects is unknown, although comparisons of larval and adult responses suggest that they might be correlated in some instances and uncorrelated in others (e.g. Du et al., 1995; Renwick et al., 2001; Renwick, 2002).

We have addressed these issues by estimating quantitative genetic variation in responses of the Colorado potato beetle (CPB), L. decemlineata (Say) (Coleoptera: Chrysomelidae), to five plant species. We used a half-sib design, and measured both larval and adult feeding consumption of each plant, as well as several facets of performance, including efficiency of neonate (newly hatched) larvae, larval weight gain, survival, development time, and adult weight at eclosion. Efficiency is a measure of growth while holding consumption constant in an analysis of covariance (ancova) framework (Raubenheimer & Simpson, 1992; Horton & Redak, 1993). We were interested chiefly in genetic correlations among traits. Across host–plant species, negative correlations (especially in performance-related characters) might imply trade-offs in fitness, a possible basis for the evolution and maintenance of specialization. Conversely, positive correlations across plants could imply not only that generalization (polyphagy) might be possible, but also that it might evolve by adding suites of encountered plant species to the diet rather than by independent adaptation to each of a number of plants. Within host–plant species, some of the genetic correlations among traits address the question of the dimensionality of adaptation. In our experiment, correlations between larval feeding and adult feeding (behavioural) responses and between larval conversion efficiency and adult feeding are germane to this question.

Because it is a major pest of potato, L. decemlineata has been studied extensively, especially with respect to behaviour (e.g. Tower, 1906; Hsiao, 1985, 1988; Hare, 1990; Jermy, 1994; Mitchell, 1994). Ancestrally, the species was distributed in Mexico and parts of the western USA, and was associated with a few (possibly only three) host species. It colonized potato in the 1830s or 1840s, and subsequently spread throughout much of North America and Eurasia, where it has been recorded on about 20 hosts, all in the Solanaceae (Tower, 1906; Casagrande, 1985; Hsiao, 1986; Jacques, 1988; Hare, 1990). We measured responses to five host species. Solanum rostratum Dunal. (buffalo bur) is believed to be the ancestral host in the USA and parts of Mexico, where the closely related species S. angustifolium Miller (not included in experiments) is also used; S. elaeagnifolium Cav. (silver-leafed nightshade) is the major host in southern Arizona and parts of Mexico. There exists geographic variation in adaptation of the beetle to at least one of these hosts, S. elaeagnifolium (Hsiao, 1978, 1985). Potato, S. tuberosum L., is ‘the most suitable host’ for USA populations of CPB (Hare, 1990), although it evokes little feeding or oviposition by populations associated with S. elaeagnifolium and S. angustifolium in Mexico and Arizona (Hsiao, 1978, 1985; Harrison & Mitchell, 1988; Lu & Logan, 1994a,b; A.G. Ehmer, personal observations). We also included in our study S. carolinense (horsenettle), to which some populations in south-eastern North America have adapted (Hare & Kennedy, 1986; Hare, 1990), and S. dulcamara L. (bitter or climbing nightshade), a European exotic that is utilized (and is suitable) only seasonally (Hare, 1983).

Materials and methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References
  9. Supporting Information

Insects and plants

Over the course of 3 days in early June 2005, 53 individual egg masses were collected from about 450 adult beetles taken from agricultural potato fields in Brookhaven, Suffolk Co., New York, USA. An average of 8.2 larvae was reared from each egg mass to pupation on S. tuberosum. As adults emerged, matings were arranged, with two dams chosen randomly for each sire (no siblings were included in any trio). A total of 33 sires were successfully mated to two dams each.

Plants were grown in a greenhouse from the following sources: S. tuberosum tubers, Superior variety, from Agway, Port Jefferson Station, NY, USA; S. rostratum seeds from Valley Seed Source, Fresno, CA, USA; S. dulcamara rootstock collected on the campus of the State University of New York, Stony Brook, NY, USA; S. carolinense seeds collected at Blandy Experimental Farm, Boyce, VA, USA; and S. elaeagnifolium seeds collected in Roseville, CA, USA. All plants were grown in standard greenhouse medium under identical conditions of light, temperature and monthly fertilization. The S. dulcamara plants grown from roots were used in the 48-h larval trials described below, while leaves harvested from the plants in the same location were used to rear larvae and assess adult consumption.

Experimental design

We conducted three main experiments: (1) a trial of individual neonate larvae in which consumption and weight gain were recorded after 48 h of feeding on the different hosts; (2) larval performance throughout development, measuring survival to adult, days to adult, and fresh adult weight; and (3) tests of host acceptance by adult beetles in sequential, no-choice feeding assays.

Measurements

Newly hatched larvae were isolated for experiments before they moved off the egg masses in search of foliage. From each dam, three larvae were assigned to each of the five plant treatments. Larvae were confined individually in 1.5-mL plastic microcentrifuge tubes (Falcon brand) with 1-cm-diameter leaf discs (punched from fresh leaves with a cork hole punch). Larvae and leaf discs were weighed on a Mettler Toledo microbalance at the start of the experiment to the nearest 0.001 μg. Leaf discs were wedged into tubes such that both upper and lower surfaces were accessible to the larvae. Tubes were opened briefly at 24 h to allow for air circulation.

After 48 h, larvae were weighed a second time and leaf discs were taped to white paper and immediately scanned in a Hewlett Packard Scanjet 3970 flat-bed scanner (Hewlett Packard, Palo Alto, CA, USA). The area of leaf disc remaining was measured from the scanned images using imagej v.1.34s (Wayne Rasband, National Institutes of Health, Bethesda, MD, USA). Mass of leaf consumed was calculated as the initial minus final weight; final weight was calculated as the final area of leaf matter remaining multiplied by the ratio of initial weight to initial area. Because wet weight of leaf discs was measured, we estimated ratios of wet to dry leaf weight for the five species by sampling control leaves throughout the course of the experiment, and used these estimates, in conjunction with a ratio of wet/dry first-instar larval weights, in order to compare efficiency across hosts in terms of dry weight. We estimated the ratio of wet/dry first instar larval weights by separately rearing and weighing 30 larvae both immediately after 48 h and after desiccation in a drying oven. The slope of the relationship of dry to wet weight was 0.156, and was highly significant (N = 30, R2 = 0.94, F1,28 = 445.8, P < 0.0001). The slope does not appear to be affected by rearing host (A.G. Ehmer, unpublished data), and a nearly identical wet/dry slope has been obtained with a different Leptinotarsa species (M.L. Forister, unpublished data).

After the 48-h trials, larvae were moved individually into large Petri dishes (lined with Kimtech tissues) and were provided daily with fresh leaves (with the cut petioles enveloped in moist cotton) of the same plant to which they had already been exposed. When larvae finished feeding, they were moved into small containers with potting soil, into which they burrowed before pupating. Adults were sexed and weighed within 24 h after eclosion (fresh adult weight). The number of days from hatching to emergence as adults was also noted. In addition, we reared six larvae from each of the 66 full sib families to pupation on S. tuberosum; of these, 271 adult beetles were sexed and tested for acceptance of the five different hosts. One-centimetre-diameter leaf discs were presented sequentially and in random order to adults confined in small Petri dishes, lined with moistened paper towelling, for periods of 30 min: long enough for beetles to find and either accept or reject the disc, but not enough time for most beetles to completely consume a disc. Following each 30-min trial, beetles were moved into clean Petri dishes for 1 h of starvation before the next trial. These trials were conducted in a Percival growth chamber with fluorescent lighting at 25.6 °C. Following each 30-min feeding trial, leaf disc remains were taped to paper and scanned for measurement in imageJ v.1.34s. The amount of leaf eaten was calculated from the area of the original leaf disc that remained.

Analyses: anova models

Variation was analysed by analysis of variance (anova), employing restricted maximum likelihood (REML), using proc mixed in sas v.9 (SAS Institute, Cary, NC, USA). Homogeneity of variance and normality for all data were evaluated using jmp v.5.0.1a (SAS Institute). Except for angular transformation of the survival fraction, no transformations were found necessary for homogeneity or normality. The significance of random factors (usually, in our data, sire and dam nested within sire) in the REML framework is estimated by comparing the ‘fit’ of a model containing the factor of interest to a model without that factor (Fry, 2004). After the significance of random factors had been evaluated, F-ratios and P-values were generated for fixed factors and interactions between fixed factors, using models containing all significant random factors. (Residuals are similarly reported from the models containing fixed factors and significant random factors.) As covariates, we used initial weight in analysing consumption and final weight in the 48-h trials, consumption in analysing efficiency (Raubenheimer & Simpson, 1992; Horton & Redak, 1993), and 48-h larval weight in analysing adult weights, survival and days to pupation. Analyses of data on adult beetles included sex as a fixed factor.

Data from each experiment were analysed first with full models involving all hosts and all possible interactions, and then with reduced models for each host. For brevity, we present only the latter results, commenting on the full models where relevant. Variance associated with sires and dams (Vs and Vd, respectively) in these host-specific analyses was determined using models that contained simple effects (e.g. covariate, sire and dam). Two-way interactions (between the covariate and sire or dam) were rarely significant (details from interactions are given in). The observational components of genetic variance estimated on each host, Vs and Vd, were converted to causal components, Vadditive (Va) and Vmaternal (Vm) with the standard calculations based on genetic relationships among siblings (Falconer, 1989): Va = 4Vs and Vm = Vd– 1/4(Va). Finally, we calculated coefficients of genetic variation, the most appropriate, mean-standardized basis for comparing variance components across traits that differ in effect size. Following Houle (1992), we used CV = 100 (inline image/inline image), where V is either Va or Vm and inline image is the mean value of the trait.

Analyses: correlations

Cross-environment correlations can be calculated with anova-based methods (Lynch & Walsh, 1998) or as simple product–moment correlations of means of half-sib families (for which P-values are more readily generated; but see Fry, 2004). Astles et al. (2005) have pointed out that if the goal is to test the null hypothesis of genetic independence, the simple product–moment correlations, although less accurate, are sufficient, because the test is conservative: half-sib family means appear to underestimate true genetic correlations. Our correlations calculated from separate anovas sometimes exceeded one (Hill & Thompson, 1978), and REML models involving two hosts (using code from Fry, 2004) often failed to converge, and so could not produce correlations. Consequently, we calculated correlations of half-sib family means based not on raw data, because our experiments were designed in an ancova framework, but instead on the residuals from general linear models that included covariates and other secondary factors (such as sex). For example, for 48-h weight gain we took the average of residuals of half-sib groups from a model that included final weight as the response and initial weight as the covariate. These analyses are based on a sample size of 33 half-sib families, with two exceptions due to mortality: for adult weight and development time, there were only 28 half-sib families raised successfully on S. rostratum, and 27 on S. carolinense.

The family means that were obtained in this way (using residuals to account for variation associated with covariates and sex) are essentially identical to breeding values estimated as best linear unbiased predictors (BLUPs) of half-sib families (Littell et al., 1996). We report both half-sib family means and BLUPs, but we focus on the former since the matrix of correlations of half-sib family means is complete (as implemented in the SAS language, BLUPs are not always able to estimate breeding values when sire-associated variation is low). Residuals for half-sib family means and BLUPs were estimated with sas v.9, and correlations were done in jmp v.5.0.1a.

We did not analyse correlations between 48-h consumption and 48-h efficiency, because the latter trait was calculated as residuals from a linear model in which weight gain was the response and consumption was the independent variable; thus efficiency should be statistically unrelated to consumption. Likewise, the correlation between efficiency and weight gain was not examined because it is biased to be a positive relationship by the way in which efficiency was calculated (as the residuals from a model with weight gain as the response variable).

Finally, bootstrap analysis (1000 replicates) was used to generate 95% confidence intervals for all correlations in simstat v.2.5.3 (the half-sib family value was the resampled unit; Provalis Research, Montreal, QC, Canada). Confidence intervals make it possible to test not only the null hypothesis that a given correlation coefficient is different from zero (which can be done without confidence intervals), but also that a coefficient is different from |1.0| (Via & Lande, 1985; Windig, 1997).

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References
  9. Supporting Information

In most respects, S. tuberosum, S. rostratum and S. dulcamara were the superior hosts for growth and survival (Table 1; anovas used to generate these comparisons are not shown, as they are lengthy and not relevant to our primary questions). S. elaeagnifolium supported relatively little growth or survival, and elicited little consumption by either adults or larvae. S. elaeagnifolium has been left out of all larval analyses beyond the 48-h trials, because only 14 individuals survived to adulthood on this plant. These differences in performance are, of course, derived from laboratory conditions, and performance in the field may or may not be comparable (Wade, 1994).

Table 1.   Main (host) effects from all experiments. In tables and figures, the plant species Solanum tuberosum, S. rostratum, S. dulcamara, S. carolinense, and S. elaeagnifolium are denoted as SoTu, SoRo, SoDu, SoCa, and SoEl, respectively. Least squares means (LS means) were generated with models including all simple effects and any significant compound effects from analyses of full models (complete results from full models not shown, see text for details).
  (A) 48 h consumption, LS means, mg (±SE) (B) 48 h weight gain, LS means, mg (±SE)(C) 48 h efficiency, wet, LS means, mg (±SE)(D) 48 h efficiency, dry, LS means, mg (±SE)(E) Adult weight, LS means, mg (±SE)(F) Days to adult, LS means, days (±SE)(G) Survival, LS means, % (±SE) (H) Adult consumption, LS means, cm2 (±SE)
  1. Lowercase letters following LS means indicate differences significant at P < 0.05 with Tukey's multiple comparisons. Efficiency values (C and D) are to be interpreted as the average weight (±SE) of larvae reared on the five hosts, standardized for amount eaten by an analysis of covariance. Values not shown for S. elaeagnifolium for E, F, and G due to low survival.

SoTu3.44 (0.092) a1.28 (0.074) a1.17 (0.080) a0.26 (0.012) a108 (1.4) a23.6 (0.17) a96.2 (5.4) a0.379 (0.011) a
SoRo2.76 (0.091) b1.63 (0.074) b1.58 (0.078) b0.28 (0.010) a102 (2.0) ab23.7 (0.27) a41.3 (5.4) b0.488 (0.011) b
SoDu2.84 (0.091) b1.16 (0.074) a1.15 (0.078) a0.21 (0.010) b99.7 (1.3) b25.7 (0.17) b99.3 (5.4) a0.312 (0.011) c
SoCa2.94 (0.092) b1.21 (0.074) a1.06 (0.079) a0.19 (0.011) bc91.4 (2.2) c26.9 (0.28) c32.0 (5.4) b0.382 (0.011) a
SoEl0.85 (0.091) c0.61 (0.074) c1.03 (0.10) a0.16 (0.012) cNANANA0.226 (0.011) d

Initial weight had a greater impact on performance and consumption on some plant species than on others (details from ancova not shown). (Following tests for homogeneity of slope, covariates were therefore dropped from models to generate the simple cross-host comparisons in Table 1.) There was also a significant interaction between amount consumed (the covariate) and host in the analysis of 48-h efficiency. However, we have chosen to ignore that interaction (which was significant at P = 0.03) rather than drop the covariate, for the purpose of comparing efficiency across hosts using ancova (Table 1c). Growth efficiency was greatest on S. rostratum, the ancestral host, and S. tuberosum, the current host, when dry weights were analysed, and was greatest for S. rostratum when wet weights were used (Table 1d).

Additive genetic variation (expressed as significant sire effects) was found for some traits, and significant dam effects were found for many, particularly larval, traits. The results from REML analyses of variance are summarized in Fig. 1 in the form of coefficients of genetic variation; complete anova tables are presented in. Sire effects were never significant on S. tuberosum or S. rostratum, and in many cases the REML models were unable to estimate nonzero sire covariance components (even nonsignificant ones) on these hosts. Four of the six significant sire effects were found for consumption (larval and adult) of the same two hosts, S. dulcamara and S. elaeagnifolium. A significant sire effect was detected for efficiency on S. carolinense. This can be interpreted as a post-ingestive effect, since the analysis examined weight gain while controlling for amount consumed across families (there was also a significant sire effect for survival on this host, see). Dam effects were pervasive in the neonate trials, and only rarely significant for other characters.

image

Figure 1.  Summary of results from REML analyses, showing coefficients of additive and maternal genetic variation for the traits studied. See Table 1 for host–plant abbreviations. Asterisks above dark bars indicate coefficients of additive variation associated with significant sire effects in REML analyses (*P < 0.05; **P < 0.01); for details of analyses see. For the two largest, but still nonsignificant sire effects, P-values are shown (adult weight and days to adult on Solanum carolinense). The significance of maternal genetic variation is not indicated here, but there were significant dam effects for all 48 h traits on all plants, as well as for adult weight on S. dulcamara and adult consumption of S. carolinense. One trait is not shown here, survival, for which there was a significant sire effect on S. carolinense (dam effects were not evaluated for survival, as each full-sib family was represented by only one value, the fraction of larvae surviving).

Download figure to PowerPoint

Additive genetic correlations can be significant and informative even if main sire effects are not significant (Fry, 1992). In general, genetic correlations were either positive or nonsignificant, and in no instance did bootstrapped confidence intervals overlap |1| (Figs 2 and 3). (See also, which indicate the few differences between Pearson product–moment and Spearman rank correlations, 95% bootstrapped confidence intervals, and the results from correlations of BLUP breeding values.)

image

Figure 2.  Genetic correlations across plant species for the seven traits studied, based on Pearson product–moment correlations. The top row is early larval traits from 48 h trials, the middle row is larval traits representing performance throughout development, and the final circle shows correlations among adult consumption.

Download figure to PowerPoint

image

Figure 3.  Genetic correlations among traits within each plant species, based on Pearson product–moment correlations. Trait abbreviations as follows: WTL, larval weight after 48 h trials; CSL, larval consumption during 48 h trials; EF, larval efficiency during 48 h trials; WTA, fresh adult weight of newly emerged adults; DY, the number of days from hatching to emergence as adult; SV, survival to emergence as adult; CSA, adult consumption in sequential host tests. Correlations were not examined between 48-h consumption and 48-h efficiency, nor between 48-h weight gain and 48-h efficiency (see Materials and methods). The matrix of correlations is incomplete for Solanum elaeagnifolium because survival past the 48 h trails was low.

Download figure to PowerPoint

The strongest additive genetic cross–host correlations were among early larval traits, especially consumption and, somewhat more weakly, 48-h weight and efficiency (Fig. 2). The distribution of correlation coefficients shows a higher mean for early larval traits than for later larval and adult traits (Fig. 4a and b).

image

Figure 4.  Histograms illustrate the distribution of Pearson correlation coefficients for three groups of traits. The y-axis indicates the number of pairwise correlations of a given magnitude, with dark bars representing correlations significant at P < 0.05. (a) Correlations across host plants for early larval traits (the top row in Fig. 2); (b) correlations across host plants for later larval and adult traits (the middle row plus adult consumption from Fig. 2); (c) correlations among traits within plant species (all the correlations expressed on the different plant species are shown here, see Fig. 3).

Download figure to PowerPoint

Within hosts, only a few traits were significantly correlated, and these may have a simple causal or developmental link (Fig. 3; see also Fig. 4c). Weight gain and 48-h consumption, for example, are positively correlated on each host. Adult weight and development time are significantly negatively correlated on three hosts, suggesting that genotypes that develop quickly also grow efficiently, getting larger in a shorter amount of time. Juvenile and adult traits (adult consumption) do not appear to be correlated, except for 48-h larval weight gain and adult consumption on S. rostratum (r = 0.48, P = 0.005), and 48-h larval efficiency and adult consumption on S. elaeagnifolium (r = 0.56, P < 0.001). Since our results have not been Bonferroni-adjusted, we are not sure how to interpret these isolated correlations. The appropriateness of Bonferroni corrections in quantitative genetic correlations is hard to assess (Lynch & Walsh, 1998), since one does not know the number of independent comparisons being made without understanding the genetic architecture involved (which is in fact the object of study) (Moran, 2003). Note that for no correlation among traits do the bootstrap-based 95% confidence intervals embrace either 1.0 or −1.0.

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References
  9. Supporting Information

The strengths of our study, relative to many quantitative genetic studies of host–plant use by phytophagous insects, are that we used a half-sib (rather than full-sib) design, enabling us to distinguish additive genetic from other (e.g. maternal-effect) sources of variation; that we scored responses to multiple plant species, enabling us to address questions related to the evolution of diet breadth (see below); and that we attempted to separate the contributions of consumption and physiological efficiency to ‘performance’ measures such as weight gain. However, the cost of including five plant species and multiple traits in our design was that we could study fewer half-sib families. Thus, a drawback of this study (shared with many others) is that our sample size (33 half-sib families) makes the detection of weak genetic correlations difficult. Statistical power with 33 observations drops below 0.8 for correlations between 0.48 and −0.48 (for a two-sided test and significance measured at P < 0.05) (Phillips, 1998). Given our interest in genetic constraints and the dimensionality of adaptation, the inability to detect such weak correlations is, however, relatively unimportant. The theoretical expectation is that only strong correlations (near |1.0|) are likely to act as significant constraints on selection (Via & Lande, 1985; Lynch & Walsh, 1998), and we assume that, likewise, genetic correlations among characters must be strong if they are to effectively reduce the dimensionality of phenotypic variation and thus enhance the response to selection.

In the following discussion, we refer to both S. tuberosum and S. rostratum as ‘major hosts’ and to S. carolinense, S. dulcamara and S. elaeagnifolium as ‘minor hosts’. All the characters we measured displayed mean differences among at least some host plants (Table 1). Solanum elaeagnifolium was an inferior food plant in almost every respect. The two other minor hosts were inferior to the two major hosts in some, but not all, respects; for example, both efficiency and either larval or adult consumption were reduced on S. carolinense and S. dulcamara. Thus, barriers to effective host use may exist at both behavioural and post-ingestive physiological levels, and could impose selection on both kinds of characters. Latheef & Harcourt (1972) showed that CPB has lower conversion efficiency on tomato, an occasional local host, than on potato.

A paucity of genetic variation in one or more of the traits we measured might hamper adaptation to certain plants. Futuyma et al. (1995), who found genetic variation in the capacity of species of Ophraella leaf beetles to use some novel plants but not others, suggested that such differences may have guided the evolutionary history of host–plant associations in this and perhaps other groups of insects. Absence of genetic variation in ecologically important traits has been described for some other species, as well (Blows & Hoffmann, 2005). We found evidence of additive genetic variance in several characters, especially larval and adult consumption, on some hosts, but not others (Fig. 1). Interestingly, additive variance was evident only on some of the ‘minor’ hosts, not on the ‘major’ hosts. A similar pattern, of greater genetic variance of host-use traits on novel or minor hosts than on normal or major hosts, has been reported both for CPB (Hare & Kennedy, 1986) and other species of insects (e.g. Tucićet al., 1997; Ueno et al., 2001). A likely explanation is that directional selection on a normal host has reduced genetic variation in traits that affect fitness on that host, whereas traits that affect fitness on minor or novel hosts are subject to lesser or no selection. Alternatively, genetic variation might be more likely to be expressed and/or detected in stressful environments, such as on a minor host where performance and survival are lower (Hoffmann & Parsons, 1991).

Negative genetic correlations in performance of insects on different plant species are often sought as evidence of trade-offs that may impose selection for specialized host preferences (Jaenike, 1990; Ballabeni et al., 2003; Milanović & Gliksman, 2004; Futuyma, in press). Joshi & Thompson (1995) hypothesized that correlations between normal and novel hosts may often be nonsignificant, but would evolve to become negative in populations that become adapted to multiple hosts. Although some negative correlations, taken as evidence of trade-offs, have been found (e.g. MacKenzie, 1996; Iriarte & Hasson, 2000; but see Tosh et al., 2004), almost all authors report nonsignificant or positive genetic correlations, even in populations with multiple hosts (Ueno et al., 2003). In analyses of full-sib family mean correlations in CPB, Hare & Kennedy (1986) found nonsignificant correlations in survival and weak (r ≤ 0.38) but significant positive correlations in adult weight on the major host S. tuberosum and the recently acquired host S. carolinense. We too find a pattern of positive or nonsignificant correlations, among both major and minor hosts (Fig. 2). This pattern is consistent with models of the evolution of specialization in which fitness on different host plants is not strongly correlated, and adaptation to a little used host is lost due to drift and fixation of disabling mutations at loci that affect performance only on that host (Fry, 1996). The higher level of additive genetic variance expressed on minor than on major host plants in our experiment also supports this view.

Positive correlations across plant species might facilitate the evolution of generalization (polyphagy). We found that many correlations in neonate consumption, weight gain, and efficiency were significant and positive (Figs 2 and 4a). Positive correlations would imply that adaptation to one plant confers ‘preadaptation’ to others, so that in principle, polyphagy might evolve by addition of suites of plants to the diet rather than independent addition of each plant species. However, adult consumption, as well as measures of ‘performance’ that encompass all of larval development (viz. survival, development time and adult weight), were not strongly correlated across host plants.

Most of the significant correlations among characters within host plants (Fig. 3) have a fairly obvious causal relationship, and so are not surprising (e.g. 48-h larval weight is correlated with consumption). Perhaps most surprisingly, on no host is consumption by larvae correlated with adult consumption, even though both characters display additive genetic variance on two of the host plants. Thus, the two behavioural characters we measured appear to be genetically independent of each other, and independent of physiological characters that affect ‘performance’.

Our data therefore provide no evidence that genetic correlations reduce the dimensionality of the phenotypic characters that may affect fitness on a particular plant. Adaptation to novel hosts may therefore entail multiple features, perhaps of such complexity as to limit the rate or direction of evolution of new host associations. This conclusion is consistent with many studies of the genetics and biology of phytophagous insects. Most of the few relevant studies of genetic correlations suggest that adult plant preferences (e.g. in oviposition) and juvenile performance are genetically independent, and many other observations imply that female oviposition preferences among plant species are not strictly concordant with larval performance (e.g. Bernays & Chapman, 1978; Fox & Lalonde, 1993; Futuyma et al., 1995; Camara, 1997; Ballabeni & Rahier, 2000; Solarz & Newman, 2001). Both the sensory physiology underlying host–plant preferences and the biochemistry of post-ingestive processing of plant toxins also are potentially complex, multidimensional characters. In L. decemlineata itself, finding and feeding on host plants are based on sex-specific, geographically variable responses of adults to plant volatiles and on contact chemoreception (Visser, 1983; Dickens, 2006), whereby feeding decisions are based on the balance between deterrent effects of compounds in a wide array of nonhost plants (Jermy, 1994) and at least two compounds that stimulate adult feeding (Müller & Renwick, 2001). Some of the steroidal alkaloids characteristic of Solanum species can have deterrent effects, whereas others do not (Hsiao, 1974, 1988; Harrison & Mitchell, 1988; Müller & Renwick, 2001). Changes in host associations may therefore entail multiple responses to chemical and perhaps other cues – responses that may well differ between larvae and adults.

Morphological traits have been shown to evolve along ‘genetic lines of least resistance’– the axis of greatest genetic variation in multivariate phenotypic space (Schluter, 1996; Begin & Roff, 2004; McGuigan et al., 2005). Adaptation that entails evolution of multiple, functionally interacting characters may often be constrained by paucity of genetic variance along the multivariate direction of selection (Blows & Hoffmann, 2005). Studies of genetic correlations do not specify the direction of selection, but they can provide some insight into the complexity of the phenotypic space in which evolution occurs (Mezey & Houle, 2005). The evolution of host–plant associations in insects can be constrained by a simple incongruity between genetic variation in ‘preference’ and ‘performance’ (Castillo-Chávez et al., 1988; Rausher, 1993). Our data suggest that these grossly defined ‘characters’ can have little genetic variation and may be genetically complex, possible contributing to the phylogenetic conservatism of host use that characterizes many clades of phytophagous insects.

Acknowledgments

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References
  9. Supporting Information

We thank M. Axelrod, D. Barnum, T. Gazi, J. Klumpp, E. Leger and L. Scarampi for their indispensable assistance in these experiments, and Hamlet Organic Garden for allowing us to collect CPB. J. Fry and J.B. Walsh kindly provided statistical advice, M.J. Wade and J.B. Walsh commented on an early draft of the manuscript, and D. Houle provided useful advice. Funding for this study was provided to D.J. Futuyma by the State University of New York at Stony Brook.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References
  9. Supporting Information
  • Astles, P.A., Moore, A.J. & Preziosi, R.F. 2005. A comparison of methods to estimate cross-environment genetic correlations. J. Evol. Biol. 19: 114122.
  • Ballabeni, P. & Rahier, M.. 2000. Performance of leaf beetle larvae on sympatric host and non-host plants. Entomol. Exp. Appl. 97: 175181.
  • Ballabeni, P., Gotthard, K., Kayumba, A. & Rahier, M. 2003. Local adaptation and ecological genetics of host–plant specialization in a leaf beetle. Oikos 101: 7078.
  • Barton, N.H. & Turelli, M. 1989. Evolutionary quantitative genetics: how little do we know? Annu. Rev. Genet. 23: 337370.
  • Begin, M. & Roff, D.A. 2004. From micro- to macroevolution through quantitative genetic variation: positive evidence from field crickets. Evolution 58: 22872304.
  • Berlocher, S.H. & Feder, J.L. 2002. Sympatric speciation in phytophagous insects: moving beyond controversy? Annu. Rev. Entomol. 47: 773815.
  • Bernays, E.A. & Chapman, R.F. 1978. Plant chemistry and acridoid feeding behavior. In: Coevolution of Plants and Animals (J. B.Harborne, ed.), pp. 100141. Academic Press, New York.
  • Bernays, E.A. & Chapman, R.F. 1994. Host–Plant Selection by Phytophagous Insects. Chapman and Hall, New York.
  • Blows, M.W. & Hoffmann, A.A. 2005. A reassessment of genetic limits to evolutionary change. Ecology 86: 13711384.
  • Blows, M.W., Chenoweth, S. & Hine, E. 2004. Orientation of the genetic variance-covariance matrix and the fitness surface for multiple male sexually-selected traits. Am. Nat. 163: 329340.
  • Bossart, J.L. 2003. Covariance of preference and performance on normal and novel hosts in a locally monophagous and locally polyphagous butterfly population. Oecologia 135: 477486.
  • Bradshaw, A.D. 1991. Genostasis and the limits to evolution. Philos. Trans. R. Soc. Lond. B, Biol. Sci. 333: 289305.
  • Brues, C.T. 1924. The specificity of food-plants in the evolution of phytophagous insects. Am. Nat. 58: 127144.
  • Camara, M.D. 1997. A recent host range expansion in Junonia coenia Hübner (Nymphalidae): oviposition preference, survival, growth, and chemical defense. Evolution 51: 873884.
  • Casagrande, R.A. 1985. The ‘‘Iowa’’ potato beetle, its discovery and spread to potato. Bull. Entomol. Soc. Am. 31: 2729.
  • Castillo-Chávez, C., Levin, S.A. & Gould, F. 1988. Physiological and behavioral adaptation to varying environments: a mathematical model. Evolution 31: 568579.
  • Dethier, V.G. 1954. Evolution of feeding preferences in phytophagous insects. Evolution 8: 3354.
  • Dickens, J.C. 2006. Plant volatiles moderate response to aggregation pheromone in Colorado potato beetle. J. App. Entomol. 130: 2631.
  • Du, Y.J., Van Loon, J.J.A. & Renwick, J.A.A. 1995. Contact chemoreception of oviposition-stimulating glucosinolates and an oviposition-deterrent cardenolide in 2 subspecies of Pieris napi. Physiol. Entomol. 20: 164174.
  • Ehrlich, P.R. & Raven, P.H. 1964. Butterflies and plants: a study in coevolution. Evolution 18: 586608.
  • Falconer, D.S. 1989. Introduction to Quantitative Genetics. John Wiley and Sons, New York.
  • Feeny, P. 1992. The evolution of chemical ecology: contributions from the study of herbivorous insects. In: Herbivores: Their Interactions With Secondary Plant Metabolites (G. A.Rosenthal & M. R.Berenbaum, eds), pp. 144. Academic Press, New York.
  • Fisher, R.A. 1930. The Genetical Theory of Natural Selection. Clarendon Press, Oxford.
  • Forister, M.L. 2005. Independent inheritance of preference and performance in hybrids between host races of Mitoura butterflies (Lepidoptera: Lycaenidae). Evolution 59: 11491155.
  • Fox, C.W. & Lalonde, R.G. 1993. Host confusion and the evolution of diet breadths. Oikos 67: 577581.
  • Fry, J.D. 1990. Trade-offs in fitness on different hosts: evidence from a selection experiment with a phytophagous mite. Am. Nat. 136: 569580.
  • Fry, J.D. 1992. The mixed-model analysis of variance applied to quantitative genetics: biological meaning of the parameters. Evolution 46: 540550.
  • Fry, J.D. 1996. The evolution of host specialization: are tradeoffs overrated? Am. Nat. 148: S84S107.
  • Fry, J.D. 2004. Estimation of genetic variances and covariances by restricted maximum likelihood using PROC MIXED. In: Genetic Analysis of Complex Traits Using SAS (A. M.Saxton, ed.), pp. 1134. SAS Institute Inc., Cary, NC.
  • Futuyma, D.J. in press. Sympatric speciation in herbivorous insects: norm or exception? In: Specialization, Speciation, and Radiation – The Evolutionary Biology of Herbivorous Insects (K.Tilmon, ed.). Oxford University Press, Oxford.
  • Futuyma, D.J. & Moreno, G. 1988. The evolution of ecological specialization. Annu. Rev. Ecol. Syst. 19: 207233.
  • Futuyma, D.J., Keese, M.C. & Funk, D.J. 1995. Genetic constraints on macroevolution: the evolution of host affiliation in the leaf beetle genus Ophraella. Evolution 49: 797809.
  • Hare, J.D. 1983. Seasonal variation in plant-insect associations: utilization of Solanum dulcamara by Leptinotarsa decemlineata. Ecology 64: 345361.
  • Hare, J.D. 1990. Ecology and management of the Colorado potato beetle. Annu. Rev. Entomol. 35: 81100.
  • Hare, J.D. & Kennedy, G.G. 1986. Genetic variation in plant-insect associations: survival of Leptinotarsa decemlineata populations on Solanum carolinense. Evolution 40: 10311043.
  • Harrison, G.D. & Mitchell, B.K. 1988. Host-plant acceptance by geographic populations of the Colorado potato beetle. J. Chem. Ecol. 14: 777778.
  • Hill, W.G. & Thompson, R. 1978. Probabilities of non-positive definite between-group or genetic covariance matrices. Biometrics 34: 429439.
  • Hoffmann, A.A. & Parsons, P.A. 1991. Evolutionary Genetics and Environmental Stress. Oxford University Press, New York.
  • Hoffmann, A.A., Hallas, R., Dean, J. & Schiffer, M. 2003. Low potential for climatic stress adaptation in a rainforest Drosophila species. Science 301: 100102.
  • Horton, D.R. & Redak, R.A. 1993. Further comments on analysis of covariance in insect dietary studies. Entomol. Exp. Appl. 69: 263275.
  • Houle, D. 1992. Comparing evolvability and variability of quantitative traits. Genetics 130: 195204.
  • Hsiao, T.H. 1974. Chemical influence on feeding behavior of Leptinotarsa beetles. In: Experimental Analysis of Insect Behavior (L.Barton Browne, ed.), pp. 237248. Springer-Verlag, Berlin.
  • Hsiao, T.H. 1978. Host plant adaptations among geographic populations of the Colorado potato beetle. Entomol. Exp. Appl. 24: 237247.
  • Hsiao, T.H. 1985. Ecophysiological and genetic aspects of geographic variations in the Colorado potato beetle. In: Proceedings of the Symposium on the Colorado Potato Beetle, XVIIth International Congress of Entomology (D. N.Ferro & R. H.Voss, eds), Mass. Agric. Exp. Stn. Res. Bull. 704, 6377.
  • Hsiao, T.H. 1986. Specificity of certain chrysomelid beetles for Solanaceae. In: Solanaceae: Biology and Systematics, Second International Symposium (W. G.D'Arcy, ed.), pp. 345363. Columbia University Press, New York.
  • Hsiao, T.H. 1988. Host specificity, seasonality and bionomics of Leptinotarsa beetles. In: Biology of Chrysomelidae (P.Jolivet, E.Petitpierrs & T. H.Hsiao, eds), pp. 581599. Kluwer, Dordrecht.
  • Iriarte, P.F. & Hasson, E. 2000. The role of the use of different host plants in the maintenance of the inversion polymorphism in the cactophilic Drosophila buzzatii. Evolution 54: 12951302.
  • Jacques, R.L.J. 1988. The Potato Beetles. E. J. Brill, New York.
  • Jaenike, J. 1990. Host specialization in phytophagous insects. Annu. Rev. Ecol. Syst. 21: 243273.
  • Jermy, T. 1994. Hypotheses on oligophagy: how far the case of the Colorado potato beetle supports them. In: Novel Aspects of the Biology of Chrysomelidae (P. H.Jolivet, M. L.Cox & E.Petitpierre, eds), pp. 127139. Kluwer Academic, Dordrecht.
  • Joshi, A. & Thompson, J.N. 1995. Trade-offs and the evolution of host specialization. Evol. Ecol. 9: 8292.
  • Keese, M.C. 1998. Performance of two monophagous leaf feeding beetles (Coleoptera: Chrysomelidae) on each other's host plant: do intrinsic factors determine host plant specialization? J. Evol. Biol. 11: 403419.
  • Latheef, M.A. & Harcourt, D.G. 1972. A quantitative study of food consumption, assimilation, and growth in Leptinotarsa decemlineata (Coleoptera: Chrysomelidae) on two host plants. Can. Entomol. 104: 12711276.
  • Littell, R.C., Milliken, G.A., Stroup, W.W. & Wolfinger, R.D.. 1996. SAS System for Mixed Models. SAS Institute, Cary, NC.
  • Lu, W. & Logan, P. 1994a. Geographic variation in larval feeding acceptance and performance of Leptinotarsa decemlineata (Coleoptera: Chrysomelidae). Ann. Entomol. Soc. Am. 87: 460469.
  • Lu, W. & Logan, P. 1994b. Genetic variation in oviposition between and within populations of Leptinotarsa decemlineata (Coleoptera: Chrysomelideae). Ann. Entomol. Soc. Am. 87: 634640.
  • Lynch, M. & Walsh, B. 1998. Genetics and the Analysis of Quantitative Traits. Sinauer Associates, Sunderland, MA.
  • MacKenzie, A. 1996. A trade-off for host plant utilization in the black bean aphid, Aphis fabae. Evolution 50: 155162.
  • McGuigan, K., Chenoweth, S.F. & Blows, M.W. 2005. Phenotypic divergence along lines of genetic variance. Am. Nat. 165: 3243.
  • Mezey, J.G. & Houle, D. 2005. The dimensionality of genetic variation for wing shape in Drosophila melanogaster. Evolution 59: 10271038.
  • Milanović, D. & Gliksman, I. 2004. Selection responses and quantitative-genetic analysis of preadult performance on two host plants in the bean weevil, Acanthoscelides obtectus. Entomol. Exp. Appl. 113: 125133.
  • Mitchell, B.K. 1994. The chemosensory basis of host–plant recognition in Chrysomelidae. In: Novel Aspects of the Biology of Chrysomelidae (P. H.Jolivet, M. L.Cox & E.Petitpierre, eds), pp. 141151. Kluwer Academic, Dordrecht.
  • Mitter, C., Farrell, B. & Futuyma, D.J. 1991. Phylogenetic studies of insect-plant interactions: insights into the genesis of diversity. Trends Ecol. Evol. 6: 290293.
  • Moran, M.D. 2003. Arguments for rejecting the sequential Bonferroni in ecological studies. Oikos 100: 403405.
  • Müller, C. & Renwick, J.A.A. 2001. Different phagostimulants in potato foliage for Manduca sexta and Leptinotarsa decemlineata. Chemoecology 11: 3741.
  • Ng, D. 1988. A novel level of interactions in plant–insect systems. Nature 334: 611613.
  • Orr, H.A. 2000. Adaptation and the cost of complexity. Evolution 54: 1320.
  • Phillips, P.C. 1998. Designing experiments to maximize the power of detecting correlations. Evolution 52: 251255.
  • Raubenheimer, D. & Simpson, S.J. 1992. Analysis of covariance: an alternative to nutritional indices. Entomol. Exp. Appl. 62: 221231.
  • Rausher, M.D. 1993. The evolution of habitat preference: avoidance and adaptation. In: Evolution of Insect Pests: Patterns of Variation (K. C.Kim & B. A.McPheron, eds), pp. 259283. Wiley, New York.
  • Renwick, J.A.A. 2002. The chemical world of crucivores: lures, treats and traps. Entomol. Exp. Appl. 104: 3542.
  • Renwick, J.A.A., Zhang, W.Q., Haribal, M., Attygalle, A.B. & Lopez, K.D. 2001. Dual chemical barriers protect a plant against different larval stages of an insect. J. Chem. Ecol. 27: 15751583.
  • Schluter, D. 1996. Adaptive radiation along genetic lines of least resistance. Evolution 50: 17661774.
  • Schluter, D. 2000. The Ecology of Adaptive Radiations. Oxford University Press, Oxford.
  • Schoonhoven, L.M., Van Loon, J.J.A. & Dicke, M. 2005. Insect–Plant Biology. Oxford University Press, Oxford.
  • Singer, M.C., Ng, D. & Thomas, C.D. 1988. Heritability of oviposition preference and its relationship to offspring performance within a single insect population. Evolution 42: 977985.
  • Solarz, S.L. & Newman, R.M. 2001. Variation in hostplant preference and performance by the milfoil weevil, Euhrychiopsis lecontei Dietz, exposed to native and exotic watermilfoils. Oecologia 126: 6672.
  • Thompson, J.N. 1988. Evolutionary ecology of the relationship between oviposition preference and performance of offspring in phytophagous insects. Entomol. Exp. Appl. 47: 134.
  • Thompson, J.N. 1996. Trade-offs in larval performance on normal and novel hosts. Entomol. Exp. Appl. 80: 133139.
  • Thompson, J.N. & Pellmyr, O. 1991. Evolution of oviposition behavior and host preference in Lepidoptera. Annu. Rev. Entomol. 36: 6589.
  • Tilmon, K.J., Wood, T.K. & Pesek, J.D. 1998. Genetic variation in performance traits and the potential for host shifts in Enchenopa treehoppers (Homoptera: Membracidae). Ann. Entomol. Soc. Am. 91: 397403.
  • Tosh, C.R., Morgan, D., Walters, K.F.A. & Douglas, A.E. 2004. The significance of overlapping plant range to a putative adaptive trade-off in the black bean aphid Aphis fabae Scop. Ecol. Entomol. 29: 488497.
  • Tower, W.D. 1906. An Investigation of Evolution in Chrysomelid Beetles of the Genus Leptinotarsa. Carnegie Institute of Washington, Washington, DC.
  • Tucić, N., Mikuljanac, S. & Stojković, O. 1997. Genetic variation and covariation among life history traits in populations of Acanthoscelides obtectus maintained on different hosts. Entomol. Exp. App. 85: 247256.
  • Ueno, H., Hasegawa, Y., Fujiyama, N. & Katakura, H. 2001. Comparison of genetic variation in growth performance on normal and novel host plants in a local population of a herbivorous ladybird beetle, Epilachna vigintioctomaculata. Heredity 87: 17.
  • Ueno, H., Fujiyama, N., Yao, I., Sato, Y. & Katakura, H. 2003. Genetic architecture for normal and novel host–plant use in two local populations of the herbivorous ladybird beetle, Epilachna pustulosa. J. Evol. Biol. 16: 883895.
  • Via, S. & Lande, R. 1985. Genotype–environment interactions and the evolution of phenotypic plasticity. Evolution 39: 505522.
  • Visser, J.H. 1983. Differential sensory perceptions of plant compounds by insects. In: Plant Resistance to Insects. ACS Symposium Series 208 (P. A.Hedino, ed.), pp. 215230. American Chemical Society, Washington, DC.
  • Wade, M.J. 1994. The biology of the imported willow leaf beetle, Plagiodera versicolora (Laicharting). In: Novel Aspects of the Biology of the Chrysomelidae (P. H.Jolivet, M. L.Cox & E.Petitpierre, eds), pp. 541547. Kluwer Academic, Dordrecht.
  • Wagner, G.P. 1988. The influence of variation and of developmental constraints on the rate of multivariate phenotypic evolution. J. Evol. Biol. 1: 4566.
  • Wagner, G.P. & Altenberg, L. 1996. Complex adaptations and evolution of evolvability. Evolution 50: 967976.
  • Wiens, J.J. & Graham, C.H. 2005. Niche conservatism: integrating evolution, ecology, and conservation biology. Annu. Rev. Ecol. Evol. Syst. 36: 519539.
  • Windig, J.J. 1997. The calculation and significance testing of genetic correlations across environments. J. Evol. Biol. 10: 853874.

Supporting Information

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References
  9. Supporting Information

Table S1 Analyses of consumption, weight gain, and efficiency by host during 48 hour trials.

Table S2 Analyses of fresh adult weight, days to emergence as adult, and survival to adult.

Table S3 Analyses of consumption in consecutive preference trials of adults.

Table S4 Correlations of early larval traits across plant species from 48 hour trials.

Table S5 Correlations of later larval traits across plant species.

Table S6 Correlations of adult consumption of the five plant species.

Table S7 Correlations of traits within each host.

FilenameFormatSizeDescription
JEB1310SA1.doc58KSupporting info item
JEB1310SA2.doc56KSupporting info item
JEB1310SA3.doc45KSupporting info item
JEB1310SA4.doc38KSupporting info item
JEB1310SA5.doc34KSupporting info item
JEB1310SA6.doc32KSupporting info item
JEB1310SA7.doc45KSupporting info item

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.