Variability in plant nutrients reduces insect herbivore performance

Journal name:
Nature
Volume:
539,
Pages:
425–427
Date published:
DOI:
doi:10.1038/nature20140
Received
Accepted
Published online

The performance and population dynamics of insect herbivores depend on the nutritive and defensive traits of their host plants1. The literature on plant–herbivore interactions focuses on plant trait mean values2, 3, 4, but recent studies showing the importance of plant genetic diversity for herbivores suggest that plant trait variance may be equally important5, 6. The consequences of plant trait variance for herbivore performance, however, have been largely overlooked. Here we report an extensive assessment of the effects of within-population plant trait variance on herbivore performance using 457 performance datasets from 53 species of insect herbivores. We show that variance in plant nutritive traits substantially reduces mean herbivore performance via non-linear averaging of performance relationships that were overwhelmingly concave down. By contrast, relationships between herbivore performance and plant defence levels were typically linear, with variance in plant defence not affecting herbivore performance via non-linear averaging. Our results demonstrate that plants contribute to the suppression of herbivore populations through variable nutrient levels, not just by having low average quality as is typically thought. We propose that this phenomenon could play a key role in the suppression of herbivore populations in natural systems, and that increased nutrient heterogeneity within agricultural crops could contribute to the sustainable control of insect pests in agroecosystems.

At a glance

Figures

  1. An illustration of Jensen’s inequality.
    Figure 1: An illustration of Jensen’s inequality.

    af, The shape of the relationship between herbivore performance and a plant trait influences the consequences of trait variance for mean performance ( ) via non-linear averaging or Jensen’s inequality. The left column (a, c, e) represents plant populations in which all plants have one trait value and no variance (plants at one location on the x axis). The right column (b, d, f) represents populations with trait variance, in which half of the plants have a high trait value and half have a low value (two plants on the x axis). The trait mean, however, is the same in the constant and variable populations. With any linear function, trait value changes result in proportional changes in performance. Mean herbivore performance is therefore equal in the absence or presence of variance (no Jensen’s effect; compare a and b). With non-linear performance functions (cf), however, trait value changes do not result in proportional performance changes, and mean performance will differ in the absence and presence of trait variance. When the relationship is concave down, mean performance will be lower in the presence of trait variance (negative Jensen’s effect; compare c and d). When the relationship is concave up, mean performance will be higher in the presence of trait variance (positive Jensen’s effect; compare e and f). Doubled-headed arrows show differences in mean performance with and without plant variance.

  2. Empirical and theoretical performance curves.
    Figure 2: Empirical and theoretical performance curves.

    a, b, Growth data from empirical studies and fitted growth curves for (a) Heliothis virescens (Lepidoptera: Noctuidae) across a range of levels of various plant secondary metabolites and (b) Helicoverpa zea (Lepidoptera: Noctuidae) across a range of levels of various plant nutrients. Data are standardized to zero mean and unit standard deviation. Red curves are significantly concave down, blue curves are concave up, and magenta curves are linear. Different symbols denote different experiments. c, d, Theoretical predictions for relationships between plant traits and herbivore performance. For simplicity, d shows one curve with an intermediate maximum, but curves that asymptote at high nutrients are also possible. Both shapes are concave down and would result in negative effects of nutrient variability.

  3. The effect of variance in plant defensive and nutritive traits on herbivore growth and survival.
    Figure 3: The effect of variance in plant defensive and nutritive traits on herbivore growth and survival.

    a, b, Each point represents one herbivore species (n = 53), jittered for visibility. Diamonds and error bars show mean values and 95% confidence intervals. Growth effects are standard deviations. Survival effects are log odds ratios. cf, The empirically estimated functions that went into the analysis that yielded a and b. Red curves are significantly concave down (negative Jensen’s effect). Magenta curves are linear (no Jensen’s effect). Blue curves are significantly concave up (positive Jensen’s effect). Curves are standardized to be on the same scale.

  4. Graphical summary of database.
    Extended Data Fig. 1: Graphical summary of database.

    ac, Number of herbivore species per order (a), mobility of feeding stage (b), and host breadth (c). b, Each mobility level indicates the maximum extent at which the feeding stage of an herbivore commonly moves. For example, species in the ‘plant’ category move within plant individuals but do not typically move between plants. Species within the ‘patch’ category readily move among neighbouring plant individuals but do not typically move between patches of plants. Species in the ‘tissue’ category are restricted to a single organ within an individual plant (for example, leaf or root). Species in the ‘region’ category readily move among plant patches across entire geographic regions. c, Host breadth categories monophagous (mono), oligophagous (oligo), and polyphagous (poly) indicate that an herbivore species feeds on plant species in one genus, plant species across multiple genera within one plant family, and plant species across two or more plant families, respectively. dh, Number of herbivore performance curves per trait type (d), defence class (e), nutrient class (f), date of publication (g), and study sample size (h).

  5. Visual representation of quantitative methods.
    Extended Data Fig. 2: Visual representation of quantitative methods.

    Diagram summarizes the bootstrapping algorithm used to calculate a distribution of Jensen’s effects for each empirical dataset for herbivore growth. For more details and for differences in methods between growth and survival, see Methods and Supplementary Methods.

  6. Jensen’s effects by plant trait type (defences and nutrients) and mobility of the feeding stage.
    Extended Data Fig. 3: Jensen’s effects by plant trait type (defences and nutrients) and mobility of the feeding stage.

    Defence variance had mean effects near to zero and nutrient variability had generally negative effects regardless of the mobility of the feeding stage of the herbivore species. Species in the ‘plant’ category move within plant individuals but do not typically move between plants. Species within the ‘patch’ category readily move among neighbouring host plants but do not typically move between patches. Species within the ‘region’ category commonly move among host plant patches. Each point is one herbivore species, jittered for visibility. Diamonds and error bars show mean values and 95% confidence intervals. See Supplementary Methods for more details.

  7. Jensen’s effects by plant trait type (defences and nutrients) and host breadth.
    Extended Data Fig. 4: Jensen’s effects by plant trait type (defences and nutrients) and host breadth.

    Defence variance had mean effects near to zero and nutrient variability had generally negative effects regardless of the host breadth of the herbivore species. Oligophagous species (‘oligo’) feed on plant species in multiple genera but are restricted to one plant family. Polyphagous species (‘poly’) feed on plant species across two or more plant families. Each point represents one herbivore species, jittered for visibility. Diamonds and error bars are mean values with 95% confidence intervals. See Supplementary Methods for more details.

  8. Funnel plots for growth and survival.
    Extended Data Fig. 5: Funnel plots for growth and survival.

    a, b, The lack of a relationship between the sample size of a study and its Jensen’s effect for growth (a) or survival (b) suggests that publication bias did not have a major influence on the results. Dashed line shows zero. Solid lines show linear regressions for growth (F1,248 = 0.23, P = 0.63, R2 = 0.0) and survival (F1,203 = 1.04, P = 0.31, R2 = 0.0).

  9. Jensen’s effect for each observation by the year of publication for growth and survival.
    Extended Data Fig. 6: Jensen’s effect for each observation by the year of publication for growth and survival.

    a, b, The lack of temporal trends in Jensen’s effects for growth (a) or survival (b) suggests that publication bias did not play a major role.

References

  1. Awmack, C. S. & Leather, S. R. Host plant quality and fecundity in herbivorous insects. Annu. Rev. Entomol. 47, 817844 (2002)
  2. Adler, F. R. & Karban, R. Defended fortresses or moving targets? Another model of inducible defenses inspired by military metaphors. Am. Nat. 144, 813832 (1994)
  3. Karban, R., Agrawal, A. A. & Mangel, M. The benefits of induced defenses against herbivores. Ecology 78, 13511355 (1997)
  4. Herrera, C. M. Multiplicity in Unity: Plant Subindividual Variation & Interactions with Animals (Univ. Chicago Press, 2009)
  5. Crutsinger, G. M. et al. Plant genotypic diversity predicts community structure and governs an ecosystem process. Science 313, 966968 (2006)
  6. McArt, S. H. & Thaler, J. S. Plant genotypic diversity reduces the rate of consumer resource utilization. Proc. R. Soc. Lond. B 280, 20130639 (2013)
  7. Esquinas-Alcázar, J. Science and society: protecting crop genetic diversity for food security: political, ethical and technical challenges. Nat. Rev. Genet. 6, 946953 (2005)
  8. Díaz, S. & Cabido, M. Vive la difference: plant functional diversity matters to ecosystem processes. Trends Ecol. Evol. 16, 646655 (2001)
  9. Wood, S. A. et al. Functional traits in agriculture: agrobiodiversity and ecosystem services. Trends Ecol. Evol. 30, 531539 (2015)
  10. Whitham, T. G. in Variable Plants and Herbivores in Natural and Managed Systems (eds Denno, R. F. & McClure, M. S.) 1541 (Academic Press, 1983)
  11. Ruel, J. J. & Ayres, M. P. Jensen’s inequality predicts effects of environmental variation. Trends Ecol. Evol. 14, 361366 (1999)
  12. Bolnick, D. I. et al. Why intraspecific trait variation matters in community ecology. Trends Ecol. Evol. 26, 183192 (2011)
  13. Jensen, J. L. W. V. Sur les fonctions convexes et les inegalites entre les valeurs moyennes. Acta Math. 30, 175193 (1906)
  14. Raubenheimer, D., Lee, K. P. & Simpson, S. J. Does Bertrand’s rule apply to macronutrients? Proc. R. Soc. Lond. B 272, 24292434 (2005)
  15. Simpson, S. J. & Raubenheimer, D. A multi-level analysis of feeding behaviour: the geometry of nutritional decisions. Philos. Trans. R. Soc. B 342, 381402 (1993)
  16. Mertz, W. The essential trace elements. Science 213, 13321338 (1981)
  17. Bertrand, G. On the role of trace substances in agriculture. Eighth Int. Congr. Appl. Chem. 28, 3040 (1912)
  18. Calabrese, E. J. & Baldwin, L. A. Toxicology rethinks its central belief. Nature 421, 691692 (2003)
  19. Ali, J. G. & Agrawal, A. A. Specialist versus generalist insect herbivores and plant defense. Trends Plant Sci. 17, 293302 (2012)
  20. Andow, D. A. Vegetational diversity and arthropod population response. Annu. Rev. Entomol. 36, 561586 (1991)
  21. Zhu, Y. et al. Genetic diversity and disease control in rice. Nature 406, 718722 (2000)
  22. Carmona, D., Lajeunesse, M. J. & Johnson, M. T. J. Plant traits that predict resistance to herbivores. Funct. Ecol. 25, 358367 (2010)
  23. Curtis, P. S., Mengersen, K., Lajeunesse, M. J., Rothstein, H. R. & Stewart, G. B. in Handbook of Meta-analysis in Ecology and Evolution (eds Koricheva, J., Gurevitch, J. & Mengersen, K.) 5260 (Princeton Univ. Press, 2013)
  24. Wood, S. N. Generalized Additive Models: An Introduction with R (Chapman and Hall/CRC, 2006)
  25. R Core Team. R: A Language and Environment for Statistical Computing. http://www.R-project.org/ (2015)
  26. Schluter, D. Estimating the form of natural selection on a quantitative trait. Evolution 42, 849861 (1988)
  27. Rosenberg, M. S., Rothstein, H. R. & Gurevitch, J. in Handbook of Meta-analysis in Ecology and Evolution (eds. Koricheva, J., Gurevitch, J. & Mengersen, K.) 6171 (Princeton Univ. Press, 2013)
  28. Viechtbauer, W. Conducting meta-analyses in R with the metafor package. J. Stat. Softw. 36, 148 (2010)
  29. Bates, D., Maechler, M., Bolker, B. M. & Walker, S. lme4: Linear mixed-effects models using Eigen and S4. R package version 1.7 (2014)
  30. Koricheva, J., Gurevitch, J. & Mengersen, K. (eds) Handbook of Meta-analysis in Ecology and Evolution (Princeton Univ. Press, 2013)

Download references

Author information

Affiliations

  1. Department of Entomology and Department of Ecology and Evolutionary Biology, Cornell University, Ithaca, New York 14853, USA

    • William C. Wetzel
  2. Center for Population Biology, University of California, Davis, Davis, California 95616, USA

    • William C. Wetzel,
    • Heather M. Kharouba &
    • Moria Robinson
  3. Department of Biology, University of Ottawa, Ottawa, Ontario K1N 6N5, Canada

    • Heather M. Kharouba
  4. Department of Environmental Science and Policy, University of California, Davis, Davis, California 95616, USA

    • Marcel Holyoak
  5. Department of Entomology and Nematology, University of California, Davis, Davis, California 95616, USA

    • Richard Karban

Contributions

W.W. conceived the project. All authors contributed to the development of the question, interpreted the results, and commented on the manuscript. W.W., H.K., and M.R. collected data and assembled the database. W.W. and M.H. developed the methods. W.W. and R.K. wrote the manuscript. M.R., H.K., and W.W. made the figures.

Competing financial interests

The authors declare no competing financial interests.

Corresponding author

Correspondence to:

Reviewer Information Nature thanks M. Ayres, B. Inouye, W. Viechtbauer and the other anonymous reviewers for their contribution to the peer review of this work.

Author details

Extended data figures and tables

Extended Data Figures

  1. Extended Data Figure 1: Graphical summary of database. (275 KB)

    ac, Number of herbivore species per order (a), mobility of feeding stage (b), and host breadth (c). b, Each mobility level indicates the maximum extent at which the feeding stage of an herbivore commonly moves. For example, species in the ‘plant’ category move within plant individuals but do not typically move between plants. Species within the ‘patch’ category readily move among neighbouring plant individuals but do not typically move between patches of plants. Species in the ‘tissue’ category are restricted to a single organ within an individual plant (for example, leaf or root). Species in the ‘region’ category readily move among plant patches across entire geographic regions. c, Host breadth categories monophagous (mono), oligophagous (oligo), and polyphagous (poly) indicate that an herbivore species feeds on plant species in one genus, plant species across multiple genera within one plant family, and plant species across two or more plant families, respectively. dh, Number of herbivore performance curves per trait type (d), defence class (e), nutrient class (f), date of publication (g), and study sample size (h).

  2. Extended Data Figure 2: Visual representation of quantitative methods. (286 KB)

    Diagram summarizes the bootstrapping algorithm used to calculate a distribution of Jensen’s effects for each empirical dataset for herbivore growth. For more details and for differences in methods between growth and survival, see Methods and Supplementary Methods.

  3. Extended Data Figure 3: Jensen’s effects by plant trait type (defences and nutrients) and mobility of the feeding stage. (86 KB)

    Defence variance had mean effects near to zero and nutrient variability had generally negative effects regardless of the mobility of the feeding stage of the herbivore species. Species in the ‘plant’ category move within plant individuals but do not typically move between plants. Species within the ‘patch’ category readily move among neighbouring host plants but do not typically move between patches. Species within the ‘region’ category commonly move among host plant patches. Each point is one herbivore species, jittered for visibility. Diamonds and error bars show mean values and 95% confidence intervals. See Supplementary Methods for more details.

  4. Extended Data Figure 4: Jensen’s effects by plant trait type (defences and nutrients) and host breadth. (85 KB)

    Defence variance had mean effects near to zero and nutrient variability had generally negative effects regardless of the host breadth of the herbivore species. Oligophagous species (‘oligo’) feed on plant species in multiple genera but are restricted to one plant family. Polyphagous species (‘poly’) feed on plant species across two or more plant families. Each point represents one herbivore species, jittered for visibility. Diamonds and error bars are mean values with 95% confidence intervals. See Supplementary Methods for more details.

  5. Extended Data Figure 5: Funnel plots for growth and survival. (81 KB)

    a, b, The lack of a relationship between the sample size of a study and its Jensen’s effect for growth (a) or survival (b) suggests that publication bias did not have a major influence on the results. Dashed line shows zero. Solid lines show linear regressions for growth (F1,248 = 0.23, P = 0.63, R2 = 0.0) and survival (F1,203 = 1.04, P = 0.31, R2 = 0.0).

  6. Extended Data Figure 6: Jensen’s effect for each observation by the year of publication for growth and survival. (99 KB)

    a, b, The lack of temporal trends in Jensen’s effects for growth (a) or survival (b) suggests that publication bias did not play a major role.

Supplementary information

PDF files

  1. Supplementary Information (606 KB)

    This file contains Supplementary Methods, Supplementary Tables, a Supplementary Discussion and Supplementary References.

Additional data