Herbivory by aboveground insects impacts plant root morphological traits

Aboveground herbivory induces physiological responses, like the release of belowground chemical defense and storage of secondary metabolites, as well as physical responses in plants, like increased root biomass production. However, studies on effects of aboveground herbivory on root morphology are scarce and until now no study tested herbivory effects under natural conditions for a large set of plant species. Therefore, in a field experiment on plant–soil interactions, I investigated the effect of aboveground insect herbivory on root morphological traits of 20 grassland plant species. For 9 of the 20 species, all individuals showed shoot damage in the presence of insect herbivores, but no damage in insect herbivore exclusions. In these 9 species root biomass increased and root morphological traits changed under herbivory towards thinner roots with increased specific root surface. In contrast, the remaining species did not differ in the number of individuals damaged, root biomass nor morphological traits with herbivores present vs. absent. The fact that aboveground herbivory resulted in thinner roots with increased specific root surface area for all species in which the herbivore exclusion manipulation altered shoot damage might indicate that plants increase nutrient uptake in response to herbivory. However, more importantly, results provide empirical evidence that aboveground herbivory impacts root morphological traits of plants. As these traits are important for the occupation of soil space, uptake processes, decomposition and interactions with soil biota, results suggest that herbivory-induced changes in root morphology might be of importance for plant–soil feedbacks and plant–plant competition.


Communicated by Leland Russell.
Electronic supplementary material The online version of this article (https://doi.org/10.1007/s11258-020-01045-w) contains supplementary material, which is available to authorized users.
Aboveground consumption not only affects aboveground biomass (Crawley 1989;Maron 1998) but also induces responses in roots (see e.g., Johnson et al. 2016a). For example, physiological responses to aboveground herbivory include induction of systemic changes and release of secondary metabolites, including defense compounds (e.g., Bardgett et al. 1998;Howe and Jander 2008;Bezemer and van Dam 2005;Karolewski et al. 2010;Johnson et al. 2016a). Physical responses can include increased root biomass production (McNickle and Evans 2018; Sarquis et al. 2019). Increased root biomass production can, however, occur by producing either few thick or multiple thin roots. Furthermore, the fact that aboveground herbivory induces changes in concentrations of phytohormones like auxin and cytocinins (e.g., Johnson et al. 2016a) that besides stimulating overall root biomass production are also involved in the development of lateral roots (Lee et al. 2018;Lymperopoulos et al. 2018) suggests that aboveground herbivory might also influence root morphology.
Although a few studies have tested the effect of aboveground biomass removal on root morphological traits, using clipping (Thorne and Frank 2009), mowing (Leuschner et al. 2013), insecticide application (Pastore and Russell 2012) or short-term application of insect herbivores under artificial conditions (Tiiva et al. 2019) until now no study directly has tested herbivory effects for a large set of species under natural conditions. Therefore, within a field experiment on plant-soil interactions with 20 plant species I specifically asked: Does aboveground herbivory impact root morphological traits of plants?

Field experiment
The impact of aboveground herbivory on root morphological traits was tested within the scope of an experiment on plant-soil interactions conducted in a meadow at a field site of the University of Potsdam (N52°24 0 29.76 00 , E13°1 0 13.74 00 , Brandenburg, Germany; see Heinze et al. 2016;Heinze and Joshi 2018). The soil in this meadow consists of nutrient poor slightly sandy loam (pH: 6-7) and the plant community has high diversity ([ 60 species).
The design of this experiment has been described elsewhere (Heinze et al. 2020). Briefly, under field conditions we tested whether aboveground herbivory affects the outcome of plant-soil feedbacks (PSFs) for 20 grassland species [10 grasses and 10 forbs (including three legumes); see Fig. 1]. Therefore, growth with and without aboveground insect herbivores of these 20 plant species was tested in soils previously conditioned by themselves (i.e., home soil) as well as by the remaining 19 species (i.e., away soil) collected in the meadow. For each species there was one home soil and one away soil. The away soil was created by mixing equal portions of each of the other 19 species. Potential differences in soil nutrient availability among soils were avoided by inoculating home and away soils (10%) into an autoclaved soil:sand mixture (five times within 24 h; 20 min, 121°C). The home and away soils for every species did not differ in plantavailable (Heinze et al. 2020) nor total nutrient concentrations (see Online Resource Table S1). In this experiment plants were grown in the meadow in buried (depth: 25 cm) pots (Deepots D25L; Stuewe & Sons; USA) to enable a standardized soil volume for root growth. To test single-plant responses, one individual was planted per pot. The pots were distributed over 10 paired plots equipped with cages which were either completely covered with fly mesh (-herbivory; mesh size: 1.3 mm; Meyer; Germany) or only shaded (? herbivory; no fly mesh at the lower 50 cm) to prevent potential differences in shade or precipitation. Each 'species by soil' combination was replicated 10 times within each herbivory treatment (i.e., ? or -herbivory), resulting in 800 pots (20 species 9 2 soils 9 2 herbivory treatments 9 10 replicates). Plants were grown for 12 weeks in summer 2017.
This study tests and reports the effects of aboveground insect herbivory on plants and their root morphological traits for three randomly chosen replicates per species, soil and herbivory treatment (i.e., 240 samples) to directly link herbivory damage with changes in root morphology.

Measurements
Before harvest, damage by chewing and leaf mining insect herbivores was visually assessed without any further discrimination of feeding guilds. Grasshoppers represent the most important herbivores in this grassland (see Heinze and Joshi 2018;Heinze et al. 2020). Biomass removal was visually estimated (in percent severity; see also, e.g., Johnson et al. 2016b) at ten randomly chosen leaves per individual plant. Furthermore, the proportion of damaged leaves was determined by counting the number of damaged as well as total leaves (incidence) on each experimental plant (see Russell et al. 2010). Severity and incidence were subsequently used to assess the shoot biomass removal by aboveground insect herbivores for whole Fig. 1 Number damaged individuals (a), specific root length (SRL, b) and specific root surface area (SRSA, c) of 20 plant species grown without (white bars) and with (gray bars) aboveground insect herbivores. In a data represent sums of damaged individuals and in b and c mean ± SE (n = 6). In a asterisks above two bars indicate significant differences after Fisher's exact tests and in b and c after t-test analysis: *P \ 0.05; *P \ 0.  Smith et al. (2005). Information on whether a plant's leaves showed any damage, or not, was used to categorize the damage status of individual plants (damaged vs. undamaged) to test whether herbivore damage affected root morphological traits.
After herbivory measurements, shoots were harvested and roots were washed. To determine root length and diameter, a representative subsample of the whole root system (max diameter: 0.47 mm) of each plant was analyzed using the WinRhizo scanner-based system (Regents Instruments, Inc., Canada). Afterward, roots (i.e., the representative subsample of the whole root system and the remaining roots per individual) were dried (48 h, 80°C) and weighed to obtain root mass. The biomass of the representative subsamples was used to calculate specific root morphological traits (except average diameter, AD) according to Ryser and Lambers (1995) and Wright and Westoby (1999): specific root length (SRL; cm/ mg), specific root surface area (SRSA; cm 2 /mg) and root tissue density (RTD; mg/cm -3 ).

Statistical analysis
All analyses were performed in R version 3.1.2 (R Developmental Core Team 2014). Before analyses residuals were checked for homogeneity of variance and tested for normality.
In addition to shoot biomass removal, I used a plant's damage status (damaged vs. undamaged) as a response variable in analyses. I made this decision because plant responses to insect herbivory are very sensitive, i.e., signals from few injured cells are sufficient to induce physiological responses (Howe and Jander 2008) including changes in phytohormones that affect root development and morphology.
To test whether aboveground herbivory (i.e.,herbivory and ? herbivory) impacts the number of damaged individuals, shoot biomass removal, root biomass and root morphological traits of plants I performed ANOVAs using generalized linear mixed models (lme4 package; Bates et al. 2015). Block (i.e., replicate) was included as random factor. P-values and degrees of freedom were estimated with Type III Kenward-Roger approximation using lmerTest (Kuznetsova et al. 2017). As this experiment also tested the effects of soils (i.e., home and away) that possibly generate differences in damage by aboveground herbivores (Heinze et al. 2019), the models included the factors 'species,' 'soil' and 'herbivory' as well as their interactions. Afterward, differences in shoot biomass removal, root biomass and root morphological traits between presence and absence of herbivores were tested using two sample t-tests for every species. Differences between functional groups (i.e., grasses vs. forbs) were tested by replacing 'species' by 'functional group' in the models and using 'species' nested in 'functional group' as additional random factor.
Because the data on plant damage status followed a binomial distribution I used the glmer-function with binomial error distribution and Fisher's exact tests as post hoc tests to evaluate species-specific differences in number of damaged individuals between theand ? herbivory plots.

Results
Different soils (home vs. away) affected the amount of biomass removal by herbivorous insects, but not the number of damaged individuals, root biomass and root morphological traits tested (see Table S2). The 'Results' section therefore focuses on effects of aboveground insect herbivory and in the figures the data for the two soils are combined.

Herbivory effects on individuals of the 20 species
The presence/absence of aboveground insect herbivores affected the number of damaged individuals and the intensity of shoot biomass removal differently for the functional groups as well as for the 20 plant species (functional group/species 9 herbivory interaction: P \ 0.001 for shoot removal and number damaged plants; see Online Resource Table S2). For 8 grass and 1 forb species, all individuals were damaged and showed substantial shoot biomass removal when grown with herbivores, but suffered no shoot damage when grown without herbivores ( Fig. 1a; Online Resource Fig. S1). The remaining species showed little shoot damage (Online Resource Fig. S1). Individuals of these species either were not damaged or were equally damaged in the ? and -herbivory plots because the herbivore exclusion was ineffective for a few species (Fig. 1a).

Herbivory effects on root biomass and root morphological traits
Aboveground herbivores influenced root biomass and root morphological traits but only for those 8 grass and 1 forb species for which the herbivory exclusion manipulation significantly altered number damaged individuals and amount of herbivore damage (species 9 herbivory interaction: root biomass: F 19,160 = 4.14; P \ 0.001; SRL: F 19,160 = 3.30; P \ 0.001; SRSA: F 19,160 = 2.06; P = 0.009; Fig. 1b Table S2, Fig. S2b, c). For these species root biomass, SRL and SRSA increased, whereas AD and RTD decreased when individuals showed damage by aboveground herbivores (Fig. 1b, c and Online Resource Fig. S2). In contrast, root biomass and morphological traits for species whose individuals were either equally damaged in the ? and -herbivory plots or not damaged showed no differences (Fig. 1b, c and Online Resource Fig. S2).

Discussion
In accordance with previous studies, damaged plants in this experiment showed an increase in root biomass with aboveground herbivory (McNickle and Evans 2018; Sarquis et al. 2019). However, more importantly, aboveground herbivory increased SRL and SRSA, but only for the 8 grass and 1 forb species for which herbivory exclusion significantly reduced herbivore damage. For one species tested in this experiment, Poa pratensis, Thorne and Frank (2009) also observed an increase in SRL after aboveground biomass removal by clipping. Several studies involving other species, however, found no impact of biomass removal on root morphological traits (Pastore and Russell 2012;Leuschner et al. 2013;Tiiva et al. 2019). An increase in SRL and most likely SRSA is presumably associated with an increase in soil resource uptake efficiency (Thorne and Frank 2009). Hence, a change of root morphological traits under herbivory towards thinner roots with increased specific root surface might increase nutrient uptake to compensate for aboveground biomass loss.
That grasses responded more strongly to the herbivory manipulation compared to forbs might be due to the fact that in this system grasshoppers are the main insect herbivores (see Heinze and Joshi 2018;Heinze et al. 2020). Grasshoppers, although being generalist herbivores (Branson and Sword 2009), predominantly feed on grasses (Pfisterer et al. 2003;Franzke et al. 2010). Hence, the high preference of grasshoppers for grasses most likely caused differences in damage between grasses and forbs. However, like in other studies, in this study a few plants showed limited insect herbivore damage in the-herbivory plots, indicating that cage exclusion did not completely prevent damage by all herbivorous insects (see e.g., Palmisano and Fox 1997;Bevill et al. 1999). Nevertheless, herbivory exclusion effectively excluded grasshoppers that caused most damage in plants in the ? herbivory plots in this experiment (J. Heinze, personal observation) and therefore most probably impacted root morphological traits.
As root morphological traits are important for the occupation of soil space (Casper and Jackson 1997), uptake processes, decomposition and interactions with soil biota , they have recently gained much attention in the contexts of PSFs (see e.g., Baxendale et al. 2014;Bergmann et al. 2016;Wilschut et al. 2019) and plant-plant competition (e.g., Ravenek et al. 2016;Semchemko et al. 2018). Hence, aboveground herbivory effects on morphological root traits might be relevant for two important research fields in plant ecology.

Potential effects on PSFs and plant-plant competition
Thinner roots with a lower RTD (i.e., lower investment in dry matter; see Ryser 1996) are likely to decompose faster than roots with high diameter and RTD and thus might impact root-litter mediated PSFs Veen et al. 2019). Furthermore, root systems with a higher SRL and SRSA will also have a higher susceptibility for pathogen infection and root herbivory (Newsham et al. 1995), potentially leading to an accumulation of negative PSFs (Kulmatiski et al. 2008). In addition, a higher specific root surface enables plants to absorb more water and nutrients, which might increase competition for limiting resources between plants and soil biota (Gustafson and Casper 2004;Manning et al. 2008) and thus potentially impact the composition of soil microbial communities involved in PSFs.
Plants with larger and more finely branched root systems absorb more limiting resources (e.g., nutrients, water) and occupy more soil space and thus are more competitive, compared to plants with large diameter roots (Ravenek et al. 2016), although the outcome of competition also depends on the spatial distribution of the resources and resource-acquiring organs (Schenk 2006). However, in general, large and well-proliferated root systems are particularly important for establishing aboveground dominance (Frank et al. 2010).
Although this study reveals important findings with implications for two important fields in plant ecology there are several unresolved issues. For example, it remains unknown whether damage type (e.g., mowing/clipping, herbivory), herbivore type (e.g., ungulates, snails, insects) or feeding type (chewing, sucking, mining) differently impact root morphological traits. Furthermore, because root morphological traits inherently act together with all other root traits, such as physiological or biotic root traits , their relative importance to feedback and competition processes may be difficult to assess.
Taken together, this study shows that aboveground herbivory impacts root morphological traits of plants, and suggests that herbivory-induced changes in root morphology might be important for the outcome of PSFs and plant-plant competition.
Acknowledgements I thank Anne Schindhelm and Joana Bergmann for help and assistance with the root trait measurements, Frank Warschau, Jürgen Augustin and Doreen Schreier for help in the field, Gabriele Gehrmann for help with the soil analysis, Ronald Wille and Stefan Saumweber for technical support for the measurements of abiotic conditions and the Botanical Garden Potsdam for their cooperation. Furthermore, I would like to thank the editor and the two reviewers for helpful comments on earlier versions of the manuscript.
Author contributions JH conceived the idea, designed the study, performed the experiment, collected and analyzed the data and wrote the manuscript.
Funding Open Access funding enabled and organized by Projekt DEAL.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.