Influence of root herbivory on plant communities in heterogeneous nutrient environments

www.newphytologist.org

127

Research

Blackwell Publishing Ltd

Influence of root herbivory on plant communities in

heterogeneous nutrient environments

Glen N. Stevens and Robert H. Jones

Department of Biological Sciences, Virginia Polytechnic Institute and State University, Blacksburg, VA 24061, USA

Summary

• While plant species respond differently to nutrient patches, the forces that drivethis variability have not been extensively examined. In particular, the role of herbivoryin modifying plant–resource interactions has been largely overlooked.• We conducted a glasshouse study in which nutrient heterogeneity and rootherbivory were manipulated, and used differences in foraging among plant speciesto predict the influence of root herbivores on these species in competition. We alsotracked the influence of neighborhood composition, heterogeneity, and herbivoryon whole-pot plant biomass.• When herbivores were added to mixed-species neighborhoods,

Eupatoriumcompositifolium

, the most precise forager, was the only plant species to displaya reduction in shoot biomass. Neighborhood composition had the greatest influenceon whole-pot biomass, followed by nutrient heterogeneity; root herbivory had thesmallest influence.• These results suggest that root herbivory is a potential cost of morphological foragingin roots. Root herbivores reduced standing biomass and influenced the relativegrowth of species in mixed communities, but their effect was not strong enough atthe density examined to overwhelm the bottom-up effects of resource distribution.

Key words:

grubs, heterogeneity, nutrient patches, precision, root foragingbehavior, Scarabaeidae.

New Phytologist

(2006)

171

: 127–136

© The Authors (2006). Journal compilation ©

New Phytologist

(2006)

doi

: 10.1111/j.1469-8137.2006.01731.x

Author for correspondence:

Robert H. Jones Tel: +1540 2319514 Fax: +1540 2319307 Email: [email protected]

Received:

14 December 2005

Accepted:

12 February 2006

Introduction

The distribution of nutrients in the soil is heterogeneous, andmany studies have demonstrated plasticity in plant responseto resource-rich patches (Hodge, 2004). While many plantswill proliferate roots (i.e. forage precisely) in nutrient-rich soilpatches, the strength of this response varies widely amongspecies in a community (Einsmann

et al

., 1999; Wijesinghe

et al

., 2001; Rajaniemi & Reynolds, 2004). Plants thataggressively forage in rich patches and simultaneously reduceroot growth in poor patches can capture nutrients moreefficiently (Hutchings & de Kroon, 1994) and may gain anadvantage over their nonproliferating competitors (Hodge

et al

., 1999; Robinson

et al

., 1999; Hutchings

et al

., 2003).Despite this potential benefit, field studies have not shownthat strong foraging responses consistently lead to improved

competitive ability (Casper

et al

., 2000; Bliss

et al

., 2002).This implies that the benefits of foraging may be offset bythe costs or risks involved (Jansen

et al

., 2005; Neatrour, 2005;Stevens & Jones, 2006). Trade-offs between the costs and benefitsof precise foraging may permit the persistence of less preciseforagers in communities (Alpert & Simms, 2002).

Exposure to root herbivory may be one of the moresignificant costs of root foraging behavior. In natural systems,root herbivores are likely to respond to rich patches, as root-feeding insects can use CO

2

concentrations as a food-sourcecue (Jones & Coaker, 1977; Brown & Gange, 1990). Therefore,root herbivore densities should be greater in nutrient-richareas where roots proliferate and respiration rates are higher(Hogberg

et al

., 2001). Exclusion of root herbivores by useof chemical insecticides has resulted in significant increasesin root lifespan (Wells

et al

., 2002) and net production of

New Phytologist

(2006)

171

: 127–136



New Phytologist

(2006)

Research128

root biomass (Stevens & Jones, 2006), and significant shifts inplant community composition (Hendrix

et al

., 1988; Brown& Gange, 1990). Even so, it is unclear whether root herbivoresinfluence interactions among plant species based upon plantroot foraging behaviors.

The central objectives of this research were to determine(1) whether root herbivory can influence plant competitionand the production of plant biomass in heterogeneous soils,and (2) whether root foraging behavior influences changes incompetitive outcomes. To meet these objectives, we designeda glasshouse experiment using seedlings of three co-occurringperennials, planted two plants per pot in single- and mixed-species treatments. Single-species treatments were used toquantify the species-specific root foraging behavioral responsesin homogeneous and heterogeneous conditions. Single-species foraging results were then used to develop hypothesesfor mixed-species treatments involving manipulations of bothnutrient heterogeneity and root herbivory. We hypothesized

a priori

that, where root feeders were present, greater precisionin foraging for nutrients would lead to less growth, becausespecies that aggressively forage will concentrate roots insmall areas (by definition), making them more susceptible tothe detrimental influences of herbivory (hypothesis 1). We furtherhypothesized that root herbivory would be concentrated innutrient-rich patches (hypothesis 2) because root herbivores seekbiologically active patches within soil. Finally, we hypothesizedthat, because most plant species forage to some extent inresponse to nutrient-rich patches, the effect of root herbivoreson total pot biomass would be greater under heterogeneousthan under homogeneous conditions (hypothesis 3).

Materials and Methods

Experimental design

We chose three perennial species, one grass (

Andropogon ternarius

Michx.) and two forbs (

Solidago altissima

L. and

Eupatoriumcompositifolium

Walt.), to create competitive neighborhoods.These species are common, co-occurring early successionalperennials in the coastal plain of the south-eastern USA. Seedsof these species were collected from a single early successionalupland site at the Savannah River Site, SC in fall 2002 (seeStevens & Jones, 2006 for further description of this fieldlocation). When selecting species for the experiment fromthe pool of available plants at the local site, we avoided speciesknown to contain significant quantities of potentially toxicsecondary compounds in their roots (e.g.

Asclepias

spp.)and members of the family Fabaceae. We chose perennialspecies that were similar in terms of their adult size and overallphenology and that we believed would be similar in terms oftheir palatability to root herbivores.

The study was conducted in glasshouses located at VirginiaTech, Blacksburg, VA, USA. Seeds were germinated in ver-miculite in late February 2003, and planted into a low-nutrient

potting soil on emergence of cotyledons. In mid-March 2003,seedlings were moved to 30 cm diameter

×

28 cm deep experi-mental pots filled with construction-grade sand and inoculatedwith approx. 2 g of soil from the seed collection site to providea source of indigenous microflora. Fiberglass window mesh(1 mm grid) was inserted in the bottoms of all pots to preventthe loss of grubs (see below) and to minimize soil loss. Twoplant seedlings were planted per pot, 7.5 cm from the pot edgeon opposite sides of the pot (Fig. 1). Pots were watered everyother day over the course of the study to prevent droughtstress. Pots were subjected to natural light and ambientphotoperiod over the course of the study.

Pot treatments

Treatments were randomly assigned to pots before their estab-lishment in the glasshouse (

n

= 198 pots). Pots were plantedwith either two seedlings from a single species (78 single-speciespots) or one seedling from each of two different species(120 mixed-species pots). Single-species pots received eitherhomogeneous or heterogeneous nutrient distribution treat-ments (described in the next paragraph), and each species

×

fertility treatment combination was replicated 13 times (3species

×

2 fertility treatments

×

13 replications = 78 pots).

Fig. 1 Pot layout for glasshouse experiments (not to scale), illustrating the heterogeneous treatment design. Dashed lines indicate axes through the pot (30 cm). Arrows indicate seedlings, 15 cm apart and 7.5 cm from the pot edge. Locations of fertilized cores (F.C., 11 cm apart) and unfertilized cores (U.C., 11 cm apart) used for root sampling are shown as dashed circles. Locations of cores and seedlings were similar for all treatments.


New Phytologist

(2006)


New Phytologist

(2006)

171

: 127–136

Research 129

Mixed-species pots received either a homogeneous or a heter-ogeneous nutrient treatment as well as a root herbivore treatment(with or without grubs; described in the last paragraph of thissection). In mixed-species pots, each treatment combinationwas replicated 10 times (3 species pairs

×

2 fertility treatments

×

2 root herbivory treatments

×

10 replications = 120 pots).Nutrient distribution treatments involved addition of

5.0 g of general-purpose slow-release fertilizer (14-14-14 NPKOsmocote® Classic; Scotts Miracle-Gro, Marysville, OH, USA)to pots. In the homogeneous treatment, fertilizer was mixeduniformly into the upper 10 cm of the pot. In the heteroge-neous treatment, 2.25 g of fertilizer was mixed into each oftwo plugs of soil (3.75 cm diameter

×

10 cm depth; hereafter‘patches’), while the remaining fertilizer was broadcast overthe surface of the pot. At the same time as fertilized patcheswere created, we removed soil from similarly spaced patchesand replaced it unamended to create unfertilized patches.Patches were placed on an axis perpendicular to that of theseedling locations (Fig. 1). These procedures resulted in twotreatments with the same overall fertility, but one treatmentprovided approximately uniform nutrient supply while theother concentrated 90% of the available mineral nutrientsinto two nutrient-rich patches representing approx. 3% of thesoil volume. Previous studies have shown little or no lateralmovement of , the most mobile of the major soilnutrients, from similar size patches (Einsmann

et al

., 1999);thus, we were able to compare root growth into fertilizedand unfertilized patches at similar distances from plants, toestimate root proliferation within heterogeneous pots.

Root herbivores (white grub larvae of Coleoptera:Scarabaeidae, hereafter ‘grubs’) were collected beginning inlate March 2003 from home gardens in Montgomery County,VA. Grubs were removed from the soil by hand and stored innative soil at 5

°

C until they were added to selected pots on 4April 2003. Three grubs were added to the center of appro-priate pots; any grub that did not dig itself beneath the soilsurface within 5 min was discarded and replaced. The grubdensity used for these additions approximates those found inthe location from which plant seeds were collected (Stevens &Jones, 2006). White grubs are polyphagous (Crutchfield &Potter, 1995; Vittum

et al

., 1999; Potter & Weld, 2002), andshould be able to consume the roots of any of the species usedin our neighborhoods. However, we did not conduct experi-ments examining the species feeding preferences of the grubs.

Harvest

Pots were harvested beginning on 10 June, 6 wk after theaddition of larvae. Above-ground portions of individualplants were removed at the soil surface. Root biomass wassampled by first removing and compositing two 3.75 cmdiameter

×

15 cm deep cores of soil at the location of thefertilized and unfertilized patches (F.C. and U.C., respectively,for fertilized and unfertilized cores), giving a total of four cores

per pot (Fig. 1). Root biomass in homogeneous pots washarvested by removing and compositing two pairs of similarlyspaced patches. The remaining roots in all pots were processedas a single sample. Thus, each pot generated two above-ground samples (individual shoots) and three soil samples(e.g. in heterogeneous conditions we had samples takenfrom fertilized cores, unfertilized cores, and the remainder ofthe pot).

Roots and grubs were separated from the soil samplesby washing over a 1-mm mesh screen. During washing, werecorded the developmental status of each grub found in eachsample as well as grub locations in pots (in fertilized patches,in pot outside of patches, etc.). Washed root sampleswere moved to the laboratory and separated into < 2.0 and> 2.0 mm diameter classes mostly by eye, with an occasionalcheck using calipers. All plant parts were dried to a constantmass at 60

°

C and weighed to determine biomass.

Data analysis

Finalizing our hypotheses on competitive responses requiredranking of species by the strength of their foraging responses.To do this, we used single-species pots under homogeneousand heterogeneous conditions to compare root behaviors(i.e. precision of foraging) of each species. As previous researchin this system showed that the finest roots (i.e. those with adiameter < 2.0 mm) were the most responsive to fertilityand herbivory (Stevens & Jones, 2006), foraging responseswere calculated using only roots < 2.0 mm in diameter. Forindividual heterogeneous pots, a root foraging index (RFI)was calculated by dividing the difference between root massin the fertilized cores and the root mass in unfertilized coresby the sum of these two measures [i.e. RFI = (F.C.

−

U.C)/(F.C. + U.C.); Fig. 1]. This calculation is a modified versionof the relative fine root mass difference used by Mou

et al

.(1997) and Einsmann

et al

. (1999). In homogeneouspots, the RFI was calculated using the absolute value of thedifferences between similarly spaced cores. As calculated,this measure compares the variability between fertilized andunfertilized cores against the background variability in rootmass observed in homogeneously fertilized pots.

We analyzed root foraging responses using two-way analysisof variance (ANOVA), with species and nutrient distributionpattern (homogeneous vs heterogeneous) as main effects.Differences in root foraging precision (i.e. root foragingindex) among the three species were investigated using Tukey’smeans comparison test. These differences were used to rankspecies in terms of their propensity to forage, and to therebyspecify our hypotheses about the influence of root herbivoreson mixed-species pots.

Before testing our hypotheses, we analyzed grub recoveryrates to assess grub survival and thus the effectiveness of ourherbivory treatments. Because harvest occurred over a 2-wkperiod coinciding with the period of adult beetle emergence,

NO3−

New Phytologist

(2006)

171

: 127–136



New Phytologist

(2006)

Research130

and there were no barriers to prevent adult insects fromleaving the pots, we predicted that the date a pot was harvestedwould influence the number of insects recovered. To test thisprediction, we tracked both the number of insects recoveredin each pot and the developmental stage (i.e. larva, pupa, oradult) and used linear regression to investigate relationshipsbetween harvest date and insect recovery. In addition, we usedANOVA to investigate the potential influences of harvestdate, neighborhood composition, and nutrient heterogeneityon grub recovery.

To test our hypothesis that the strength of the root foragingresponse of a plant species (measured as RFI) would be asso-ciated with increased exposure to root herbivory (i.e. hypoth-esis 1), we compared the above-ground (shoot) biomass ofindividual species in mixed-species pots at the time of harvest.We used only shoot biomass (rather than whole-plantbiomass) to assess these differences because we were not ableto separate the roots of the different species. For each species,we analyzed the effects of the presence or absence of grubs, thespecies of competitor, the pattern of nutrient distribution,and their interactions on the shoot biomass of the targetspecies using three-way ANOVA. Least-squares means wereused in post-hoc means separation tests because of unequalreplication, as one mixed-species pot was damaged during theexperiment.

To test our hypothesis that root herbivores would selectivelyforage in nutrient-rich patches (hypothesis 2), we used dataon both grub distributions in the pots and the effects of thegrubs on core root mass. Grub distributions were analyzedusing a grub foraging index (GFI). This GFI used a similarformula to that given above for RFI, based on numbers ofgrubs rather than root biomass; in this case, GFI = (F.C.

−

U.C.)/(F.C. + U.C.). In instances where no grubs werefound in either set of cores, resulting in a zero value for thedenominator, a zero GFI value was recorded. Using thisformula, an individual pot would also have a zero GFI valuewhen the same number of grubs was found in both sets ofcores. We used nonparametric analyses to test for the separateinfluences of nutrient distribution (homogeneous and heter-ogeneous treatments) and plant neighborhood compositionon grub foraging. To analyze the effect of root herbivory onroot biomass in patches, we calculated the difference in rootbiomass between fertilized and unfertilized cores of each pot.We then tested the effect of root herbivores on this differenceusing ANOVA, with root herbivores, neighborhood composi-tion, and nutrient distribution as treatments.

To test hypothesis 3, that root herbivore effects on total potbiomass would be greater under heterogeneous conditions,we tested the individual and combined effects of nutrientheterogeneity, root herbivory, and neighborhood compositionon whole-pot (root + shoot) biomass using three-way ANOVA.All statistical comparisons were performed using StatisticalAnalysis System (SAS) software, version 9.2 (SAS InstituteInc., Cary, NC, USA).

Results

Single-species pots

The three species used in this experiment differed significantlyin their average biomass at harvest (

F

2,72

= 173,

P

< 0.0001).In single-species pots, biomass was lowest in pots plantedwith

A. ternarius

[7.46

±

0.74 g, mean

±

standard error (SE)],intermediate in the

S. altissima

pots (32.21

±

1.83 g), andgreatest for

E. compositifolium

(38.91

±

2.02 g).Regardless of species, mean whole-pot root and shoot

biomasses both increased by more than 40% in heterogeneousconditions relative to homogeneous conditions (

F

1,72

> 19.0,

P

< 0.0001 for both root and shoot biomasses). The increasein mean shoot biomass demonstrated by

E. compositifolium

(51%) was the greatest of the three species (species by nutrientdistribution interaction significant;

F

2,72

= 6.60,

P

< 0.01).The magnitudes of the root biomass responses were similaramong species.

Nutrient heterogeneity also resulted in a shift in rootbiomass distribution. Analysis of monoculture pots revealedsignificant nutrient distribution effects on our root foragingindex that were independent of species (Table 1). In heteroge-neous pots, the average difference in root biomass betweenfertilized and unfertilized cores was 55%, with biomass infertilized cores always greater than that in unfertilized cores;cores from similar locations in homogeneous pots differedby an average of only 12.6%. While each of the three speciesresponded to nutrient heterogeneity by proliferating rootsin the nutrient patches, the response of

E. compositifolium

was significantly greater than that of both

A. ternarius

and

S. altissima

(Table 1, Fig. 2). The strength of the root foragingresponse did not differ between

A. ternarius

and

S. altissima

.These results indicated that

E. compositifolium

was the mostprecise forager of the three species analyzed. We used this result topredict that

E. compositifolium

, as the most precise forager, shouldshow the greatest negative response to the presence of grubs.

Grub recovery from pots

The number of insects found in pots at harvest declined signi-ficantly over the course of harvest (regression of harvest date

Table 1 Results of analysis of variance (ANOVA) of the influence of nutrient distribution on root foraging index (RFI) for each of the three species

Source d.f. MS F P

Species 2 0.04 2.63 0.08Nutrient distribution 1 3.61 253 < 0.0001Species × nutrient distribution 2 0.09 5.99 < 0.01Error 72 0.01

d.f., degrees of freedom; MS, mean square.


New Phytologist

(2006)


New Phytologist

(2006)

171

: 127–136

Research 131

on recovery significant;

P

< 0.01,

R

2

= 0.43). The regressionequation indicated that, on the first day of harvest (date = 0),an average of 2.72

±

0.19 (mean

±

SE) insects were recoveredfrom pots. The number of grubs recovered from pots did notdiffer between neighborhood types or nutrient heterogeneitytreatments (all main effects and interactions not significant;all

F

1

or

2,48

< 0.80,

P

> 0.38).Recovery ranged from a maximum of three (100% recovery)

to zero (no insects recovered). Averaging across harvest dates,we recovered 1.65

±

0.12 (mean

±

SE) insects (i.e. grubs,larvae, or adult stages of root herbivores) from each of the potsto which grubs were added. Larvae represented the dominantdevelopmental stage found at harvest (66 of the 99 recoveredinsects), and the number of larvae found in pots also declinedsignificantly over the course of the harvest (regression,

P

< 0.05).While pupae and adults were found in the pots, they were foundless often than larvae (20 and 13 of each were recovered,respectively) and their recovery did not change through time.

Grub effects on species in competition (hypothesis 1)

As hypothesized,

E. compositifolium

was the most sensitiveof the three species to root herbivory (Fig. 3). Results of thethree-way ANOVA showed varying trends for each of thespecies (Table 2). Nutrient heterogeneity and the identityof the competitor had significant or very nearly significantinfluences on the shoot biomass of all three species, but only

E. compositifolium

responded significantly to grubs. Further,there were no significant interactions between grubs and othermain effects for any of the species.

For

A. ternarius

, the largest effect on shoot biomass wasthat of the identity of the competitor in the pot.

A. ternarius

was 43% smaller when in competition with

E. compositifolium

than when in competition with

S. altissima (average shoot

biomass 1.0 g vs 1.7 g, respectively). Although the differencewas not statistically significant, A. ternarius appeared to be largerin heterogeneous treatments than in homogeneous treatments(average shoot biomass 1.5 g vs 1.1 g, respectively). Rootherbivory effects were not detected.

The shoot biomass of E. compositifolium was significantlyinfluenced by each of the main treatment effects, andalso displayed an interaction between species of competitorand heterogeneity. Specifically, when in competition withA. ternarius, E. compositifolium biomass was 50% greater inheterogeneous than in homogeneous pots (17.2 g vs 11.3 g,respectively), but when in competition with S. altissima,E. compositifolium biomass was similar between heterogene-ous and homogeneous pots (12.1 and 11.8 g, respectively).E. compositifolium shoots were approximately 20% smaller inthe presence of grubs (Fig. 3).

Solidago altissima biomass was significantly affected bynutrient heterogeneity and the identity of the competitor inthe pot, but was not significantly affected by root herbivory(Table 2). Shoot biomass was 40% higher on average inheterogeneous pots than in homogeneous pots (10.5 vs7.4 g), and was 21% lower on average when in competitionwith E. compositifolium than when in competition withA. ternarius (7.9 vs 10.0 g).

Grub responses to heterogeneity (hypothesis 2)

Nutrient heterogeneity resulted in a statistically significantchange in the grub foraging index (nonparametric test ofheterogeneity on GFI, Kruskal–Wallis χ2 = 6.11, P = 0.01),with higher values displayed in the heterogeneous nutrienttreatment. However, although the mean GFI values for eachnutrient treatment were different from each other, neither wassignificantly different from zero. Mean GFI values (± SE)

Fig. 2 Root foraging index (RFI) calculated under homogeneous and heterogeneous conditions for each of the species (Andropogon ternarius, Eupatorium compositifolium and Solidago altissima) in this experiment. Bars show means + standard errors. Different lowercase letters indicate significantly different values at the P < 0.05 level.

Fig. 3 Shoot biomass of individual species (Andropogon ternarius, Eupatorium compositifolium and Solidago altissima) in mixed-species pots in response to grub additions. Bars show means + standard errors. The asterisk indicates a significant difference (P < 0.05) in shoot biomass of E. compositifolium between root herbivore treatments.

New Phytologist (2006) 171: 127–136 www.newphytologist.org © The Authors (2006). Journal compilation © New Phytologist (2006)

Research132

for homogeneous pots were negative (−0.077 ± 0.057), whilevalues for heterogeneous pots were positive (0.081 ± 0.052).Grub foraging did not differ between neighborhood types(Kruskal–Wallis χ2 = 2.36, P = 0.31).

The influence of grubs on the difference in standing rootbiomass between cores depended on nutrient distribution (grub× heterogeneity interaction significant; F1,107 = 5.63, P = 0.02).Under homogeneous conditions, there were no significantdifferences between cores, regardless of the presence of grubs(Fig. 4). Under heterogeneous conditions, while root biomasswas always higher in fertilized cores than in unfertilized cores,grubs reduced this difference by more than 20%.

Effects on neighborhood biomass (hypothesis 3)

Plant neighborhood biomass (i.e. whole pot root + shootbiomass) in mixed-species pots was significantly influenced by

Table 2 Results of analyses of variance (ANOVAs) of the shoot biomass responses of Andropogon ternarius, Eupatorium compositifolium, and Solidago altissima in mixed-species pots to nutrient heterogeneity, competitor identity, the presence of grubs, and the interactions between these main effects

Target species Source d.f. MS F P

A. ternarius Nutrient distribution 1 3.77 3.26 0.08Competitor identity 1 10.1 8.72 < 0.01Root herbivory 1 1.46 1.27 0.26Nutrient × competitor 1 1.34 1.16 0.28Nutrient × herbivory 1 2.47 2.13 0.15Competitor × herbivory 1 0.00 0.00 0.99Nutrient × competitor × herbivory 1 0.26 0.23 0.63Error 71 1.16

E. compositifolium Nutrient distribution 1 160 8.73 < 0.01Competitor identity 1 106 4.96 0.03Root herbivory 1 164 7.69 < 0.01Nutrient × competitor 1 186 8.73 < 0.01Nutrient × herbivory 1 10.6 0.49 0.48Competitor × herbivory 1 0.73 0.03 0.85Nutrient × competitor × herbivory 1 0.07 0.00 0.96Error 71 21.4

S. altissima Nutrient distribution 1 186 10.30 < 0.01Competitor identity 1 82.2 5.11 0.03Root herbivory 1 1.62 0.09 0.77Nutrient × competitor 1 28.2 1.56 0.22Nutrient × herbivory 1 7.56 0.42 0.52Competitor × herbivory 1 3.47 0.19 0.66Nutrient × competitor × herbivory 1 18.4 1.02 0.32Error 72 18.1

Factors highlighted in bold are significant at P < 0.05.d.f., degrees of freedom; MS, mean square.

Fig. 4 Influence of root herbivores on the difference in root mass between patches in homogeneous and heterogeneous nutrient treatments. Bars show means and standard errors. The asterisk indicates a significant difference (P < 0.05) between bars, representing a significant reduction by grubs of root mass variability in heterogeneously fertilized pots.

© The Authors (2006). Journal compilation © New Phytologist (2006) www.newphytologist.org New Phytologist (2006) 171: 127–136

Research 133

neighborhood composition, the pattern in nutrient distribution(homogeneous vs heterogeneous), and the presence of grubs(Table 3). Average neighborhood biomass was 26% greaterin heterogeneous pots than in homogeneous pots, and wasreduced by an average of nearly 11% in pots where grubs wereadded (Fig. 5). Although grub effects on neighborhood biomassappeared stronger in heterogeneous than in homogeneouspots (Fig. 6) and post-hoc means separation suggested that theseeffects may be significant (P = 0.07), there were no statisticallysignificant interactions between treatments (Table 3).

Discussion

We proposed a model of soil–root–herbivore interactionsin which root herbivores would be attracted to nutrient-richpatches, and thereby have a stronger effect on plant specieswith a greater tendency to forage for patchily distributedsoil nutrients. The observed results lend support to this model.Most importantly, root herbivory appears to be a significantcost that may influence species differentially based upon rootforaging behaviors. Plant species that aggressively forage forresources in nutrient-rich microsites may be at increased riskof root herbivory, reducing the net benefit of root foraging.

The data support a prediction from our first hypothesisthat the tendency of a plant species to forage for soil nutrients

Source d.f. MS F P

Neighborhood composition 2 1067 25.1 < 0.0001Nutrient heterogeneity 1 667 15.7 < 0.001Root herbivory 1 171 4.01 0.05Neighborhood × nutrient 2 93.9 2.21 0.12Neighborhood × herbivory 2 31.6 0.74 0.48Nutrient × herbivory 1 93.9 2.20 0.14Neighborhood × nutrient × herbivory 2 15.3 0.36 0.70Error 107 42.6

Factors highlighted in bold are significant at P < 0.05.d.f., degrees of freedom; MS, mean square.

Table 3 Results of analysis of variance (ANOVA) of whole-pot biomass response to neighborhood species composition, fertility, and root herbivory treatments, as well as interactions between these main effects

Fig. 5 Influences of neighborhood composition (a), nutrient distribution (b), and the presence of root herbivores (c) on neighborhood biomass. (A.t., Andropogon ternarius; E.c., Eupatorium compositifolium; S.a., Solidago altissima). Bars show means + standard errors. Within each panel, different lowercase letters indicate significantly different values at the P < 0.05 level.

Fig. 6 Influence of resource distribution and root herbivory on plant neighborhood biomass. Bars show means + standard errors. Different lowercase letters indicate significantly different values at the P < 0.05 level.


Research134

will be associated with reduced yield when herbivores arepresent. Eupatorium compositifolium foraged more preciselythan either A. ternarius or S. altissima, and under conditionsof interspecific competition it was the only species of thethree to show significant reductions in yield when grubswere present. However, there is a caveat to this conclusion:E. compositifolium was smaller in pots with grubs regardlessof the pattern of nutrient distribution, suggesting that anytrade-off between precision of foraging and vulnerability toherbivory does not depend strictly on local nutrient conditions.In addition, while it is plausible that we observed a causalrelationship between the precise foraging of E. compositifoliumand its vulnerability to herbivory, the results we observed mayhave resulted from root herbivore preferences for the roots ofE. compositifolium, a factor we did not examine.

Although previous studies have shown that precise foragingspecies may have a competitive advantage in mixed-speciescommunities (Robinson et al., 1999; Fransen et al., 2001;Day et al., 2003), few studies have explored potential costsof root foraging behaviors. Precise foragers benefit from anincreased ability to take up mineral resources such as nitrogenfrom resource-rich patches (Robinson et al., 1999; Hodge et al.,1999). However, if root herbivory is concentrated in nutrient-rich patches or on precisely foraging species, as suggested bythese results and related research (Stevens & Jones, 2006),root herbivory represents an unpredictable and potentiallysignificant cost of morphological root foraging. Species thatpreferentially forage in patches may be at a higher risk ofherbivory than less precise foragers. To the extent that rootherbivory constrains the benefits of root foraging, rootherbivores may allow less precise species that are presumablyless efficient at nutrient capture to persist in a community insituations where they would otherwise be outcompeted. Futureresearch considering potential costs associated with foragingmay help to explain the range of foraging responses seen evenbetween species from similar environments and with similarlife histories (Einsmann et al., 1999; Rajaniemi & Reynolds,2004).

Support for our second hypothesis, that root herbivoreswould forage preferentially in nutrient-rich patches, was mixed.While we did see differences in grub distributions betweenhomogeneous and heterogeneous treatments, this significantdifference was attributable in large part to the concentrationsof herbivores in one set of cores in the homogeneous pots. Asboth sets of cores in the homogeneous pots were theoreticallyequivalent, we conclude that the differences in foragingbetween heterogeneous and homogeneous pots result fromrandom chance. The lack of a foraging response to nutrient-rich cores was surprising, given that some root herbivoresforage in response to CO2 signals ( Jones & Coaker, 1977),which we presume were higher in fertilized patches becauseof the higher root biomass in these patches. However, whilescarab larvae are fairly mobile and do tend to aggregate, thefactors that control their behaviors are not well understood.

The design of our experiment, where nutrient patchesrepresented only 3% of the soil volume, may have contributedto the small numbers of grubs we found in cores, regardlessof fertility. Additionally, as on average more than 90% ofthe root biomass in the pots was found outside the patches,grubs would not have been forced to forage in patches tofind roots.

Although the grubs themselves were not concentratedwithin nutrient-rich patches, they significantly reduced rootproliferation in these patches. In heterogeneous pots, grubssignificantly reduced the difference in root biomass betweenfertilized and unfertilized cores (Fig. 4). Salt concentrationsin fertilizer patches may have caused grubs to avoid the innerareas of the patch, and to feed largely on the patch periphery.It is also possible that some of the grubs that fed in the patcheshad descended into deeper portions of the pots in advance ofpupation. Indeed, one-third of the grubs we recovered fromthe pots were either pupae or adults, and pupation generallytakes place deeper in the soil profile than does feeding (Vittumet al., 1999). Regardless of their final distribution, the reduc-tion in root biomass in fertilized relative to unfertilized coressupports the notion that root herbivores were attracted tonutrient-rich microsites. Similar results were seen in fieldstudies using root ingrowth cores (Stevens & Jones, 2006),where root herbivores reduced root biomass to a greater extentin fertilized cores than in unfertilized cores.

Our third hypothesis, that grub effects on neighborhoodbiomass would be stronger in heterogeneous pots than inhomogeneous pots, was not supported by the experimentalresults. The experiment involved manipulations of neighbor-hood composition, nutrient distribution, and root herbivory,and we observed significant effects of all three of these maintreatments on neighborhood biomass (Fig. 5), but no interac-tions between main effects. Interestingly, although we did notobserve the statistically significant interaction we predicted,grubs reduced whole-pot biomass by an average of 16% inheterogeneous pots, but only reduced biomass by 2.6%in homogeneous pots (Fig. 6). One source of variability thatmay have reduced our ability to resolve potential interactionswas variability in the magnitude of herbivory (i.e. root herbivorefeeding days) among pots. Only the larval stages of thesegrubs feed on roots, and, while the numbers of pupae or adultsfound in pots did not vary according to any treatment (datanot shown), their presence indicates that not all pots weresubject to identical levels of herbivory.

Ranking of plant species composition, nutrient distribution,and root herbivory in terms of their effects on plant neighbor-hood biomass suggests that plant species composition wasthe most significant driver of total biomass production overthe course of the study, as mean biomass at harvest for thedifferent neighborhood types varied by approximately 60%.Nutrient distribution was the next strongest driver, with totalpot biomass 26% greater in heterogeneous than in homogeneouspots. The effect of root herbivores was the weakest of the

© The Authors (2006). Journal compilation © New Phytologist (2006) www.newphytologist.org New Phytologist (2006) 171: 127–136

Research 135

three treatment effects, as grubs reduced biomass by approx.11% where they were present. Although these results examinedonly a small component of the plant community present at ourstudy site, at the herbivore density we examined, bottom-upeffects of neighborhood composition and nutrient supplydominated. The relative strengths of these forces are similar tothose observed in recent experiments by De Deyn et al. (2003,2004), where microcosm biomass was controlled largely bystrong bottom-up effects of resource supply, but soil fauna ledto increased plant diversity by reducing the dominance ofstrong resource competitors.

It is important to note that the trends we observed in thisstudy, where herbivore effects were weaker than the effects offertility, were most certainly influenced by the herbivoredensity we selected. Adding three grubs to each pot resultedin a density of approximately 40 grubs m−2, or four grubs persquare foot. This density is the same as that observed in thefield sites where the seeds were collected (Stevens & Jones,2006) and is similar to that reported during infestations ofwhite grubs in shortgrass and mixed-grass prairie (Ueckert,1979; Lura & Nyren, 1992). However, white grub densitiescan be much higher in turf (more than 100 per square foot;Vittum et al., 1999) and other settings. Densities of rootherbivores, as with most soil fauna, are quite heterogeneous(Ettema & Wardle, 2002), and we hypothesize that therelative influence of these fauna on plant communities varieswith root herbivore density.

Overall, these results demonstrate the potential for rootfeeders to influence plant species differentially based atleast in part on their root foraging behaviors. This supportsan emerging, complex view of nutrient–plant–herbivoreinteractions, and further supports the need to consider the roleof soil fauna in plant community interactions (Hunter, 2001;De Deyn et al., 2004).

Acknowledgements

We would like to thank Sean Moore, Kim Nguyen, DustinPierson, Julie Reimer, Julia Showalter, Ben Templeton, AbigailVitale, and Debbie Wiley for their assistance. Funding forthis project was provided by National Science FoundationGrant DEB-0308847, the Virginia Tech Graduate StudentAssembly, and the Virginia Tech Department of BiologicalSciences. The manuscript was improved by helpful commentsfrom Lynn Adler, Ed Lewis, Maury Valett, and three anonymousreviewers. We appreciate the use of the Savannah River Site,a National Environmental Research Park, and the assistanceof the USDA Forest Service.

ReferencesAlpert P, Simms EL. 2002. The relative advantages of plasticity and fixity in

different environments: when is it good for a plant to adjust? Evolutionary Ecology 16: 285–297.

Bliss KM, Jones RH, Mitchell RJ, Mou P. 2002. Are competitive

interactions influenced by spatial nutrient heterogeneity and root foraging behavior? New Phytologist 154: 409–417.

Brown VK, Gange AC. 1990. Insect herbivory below ground. Advances in Ecological Research 20: 1–58.

Casper BB, Cahill JF, Jackson RB. 2000. Plant competition in spatially heterogeneous environments. In: Hutchings MJ, John EA, Stewart AJA, eds. The ecological consequences of environmental heterogeneity. Oxford, UK: Blackwell Science, 111–130.

Crutchfield BA, Potter DA. 1995. Feeding by Japanese beetle and southern masked chafer on lawn weeds. Crop Science 35: 1681–1684.

Day KJ, John EA, Hutchings MJ. 2003. The effects of spatially heterogeneous nutrient supply on yield, intensity of competition and root placement patterns in Briza media and Festuca ovina. Functional Ecology 17: 454–463.

De Deyn GB, Raaijmakers CE, van der Putten WH. 2004. Plant community development is affected by nutrients and soil biota. Journal of Ecology 92: 824–834.

De Deyn GB, Raajimakers CE, Zoomer HR, Berg MP, de Ruiter PC, Verhoef HA, Bezemer TM, van der Putten WH. 2003. Soil invertebrate fauna enhances grassland succession and diversity. Nature 422: 711–713.

Einsmann JC, Jones RH, Pu M, Mitchell RJ. 1999. Nutrient foraging traits in 10 co-occurring plant species of contrasting life forms. Journal of Ecology 87: 609–619.

Ettema CH, Wardle DA. 2002. Spatial soil ecology. Trends in Ecology and Evolution. 17: 177–183.

Fransen B, de Kroon H, Berendse F. 2001. Soil nutrient heterogeneity alters competition between two perennial grass species. Ecology 82: 2534–2546.

Hendrix SD, Brown VK, Gange AC. 1988. Effects of insect herbivory on early plant succession – comparison of an English site and an American site. Biological Journal of the Linnean Society 35: 205–216.

Hodge A. 2004. The plastic plant: root responses to heterogeneous supplies of nutrients. New Phytologist 162: 9–24.

Hodge A, Robinson D, Griffiths BS, Fitter AH. 1999. Why plants bother: root proliferation results in increased nitrogen capture from an organic patch when two grasses compete. Plant, Cell & Environment 22: 811–820.

Hogberg P, Nordgren A, Buchmann N, Taylor AFS, Ekblad A, Hogberg MN, Nyberg G, Ottosson-Lofvenius M, Read DJ. 2001. Large-scale forest girdling shows that current photosynthesis drives soil respiration. Nature 411: 789–792.

Hunter MD. 2001. Out of sight, out of mind: the impacts of root-feeding insects in natural and managed systems. Agricultural and Forest Entomology 3: 3–9.

Hutchings MJ, de Kroon H. 1994. Foraging in plants: the role of morphological plasticity in resource acquisition. Advances in Ecological Research 25: 159–238.

Hutchings MJ, John EA, Wijesinghe DK. 2003. Toward understanding the consequences of soil heterogeneity for plant populations and communities. Ecology 84: 2322–2334.

Jansen C, van de Steeg HM, de Kroon H. 2005. Investigating a trade-off in root morphological responses to a heterogeneous nutrient supply and to flooding. Functional Ecology 19: 952–960.

Jones OT, Coaker TH. 1977. Oriented responses of carrot fly larvae, Psila rosae, to plant odors, carbon-dioxide and carrot root volatiles. Physiological Entomology 2: 189–197.

Lura CL, Nyren PE. 1992. Some effects of a white grub infestation on northern mixed-grass prairie. Journal of Range Management 45: 352–354.

Mou PP, Mitchell RJ, Jones RH. 1997. Root distribution of two tree species under a heterogeneous nutrient environment. Journal of Applied Ecology 34: 645–656.

Neatrour MA. 2005. Factors affecting root system response to nutrient heterogeneity in forested wetland ecosystems. PhD thesis, Virginia Polytechnic Institute and State University, Blacksburg, VA, USA.


Research136

Potter DA, Weld DW. 2002. Biology and management of the Japanese beetle. Annual Review of Entomology 47: 175–205.

Rajaniemi TK, Reynolds HL. 2004. Root foraging for patchy resources in eight herbaceous plant species. Oecologia 141: 519–525.

Robinson D, Hodge A, Griffiths BS, Fitter AH. 1999. Plant root proliferation in nitrogen-rich patches confers competitive advantage. Proceedings of the Royal Society of London Series B: Biological Sciences 266: 431–435.

Stevens GN, Jones RH. 2006. Patterns in soil fertility and root herbivory interact to influence fine root dynamics. Ecology 87: 616–627.

Ueckert DN. 1979. Impact of a white grub (Phyllophaga crinita) on a shortgrass community and evaluation of selected rehabilitation practices. Journal of Range Management 32: 445–448.

Vittum PJ, Villani MG, Tashiro H. 1999. Turfgrass insects of the United States and Canada, 2nd edn. Ithaca, NY, USA: Cornell University Press.

Wells CE, Glenn DM, Eissenstat DM. 2002. Soil insects alter fine root demography in peach (Prunus persica). Plant, Cell & Environment 25: 431–439.

Wijesinghe DK, John EA, Beurskens S, Hutchings MJ. 2001. Root system size and precision in nutrient foraging: responses to spatial pattern of nutrient supply in six herbaceous species. Journal of Ecology 89: 972–983.

About New Phytologist

• New Phytologist is owned by a non-profit-making charitable trust dedicated to the promotion of plant science, facilitating projectsfrom symposia to open access for our Tansley reviews. Complete information is available at www.newphytologist.org.

• Regular papers, Letters, Research reviews, Rapid reports and both Modelling/Theory and Methods papers are encouraged.We are committed to rapid processing, from online submission through to publication ‘as-ready’ via OnlineEarly – the 2004 averagesubmission to decision time was just 30 days. Online-only colour is free, and essential print colour costs will be met if necessary.We also provide 25 offprints as well as a PDF for each article.

• For online summaries and ToC alerts, go to the website and click on ‘Journal online’. You can take out a personal subscription tothe journal for a fraction of the institutional price. Rates start at £109 in Europe/$202 in the USA & Canada for the online edition(click on ‘Subscribe’ at the website).

• If you have any questions, do get in touch with Central Office ([email protected]; tel +44 1524 594691) or, for a localcontact in North America, the US Office ([email protected]; tel +1 865 576 5261).

Influence of root herbivory on plant communities in heterogeneous nutrient environments

Documents