Root competition: beyond resource depletion


H. Jochen Schenk (tel. +1 714 278 3678; fax +1 714 278 3426; e-mail


  • 1Root competition is defined as a reduction in the availability of a soil resource to roots that is caused by other roots. Resource availability to competitors can be affected through resource depletion (scramble competition) and by mechanisms that inhibit access of other roots to resources (contest competition, such as allelopathy).
  • 2It has been proposed that soil heterogeneity can cause size-asymmetric root competition. Support for this hypothesis is limited and contradictory, possibly because resource uptake is affected more by the amount and spatial distribution of resource-acquiring organs, relative to the spatial distribution of resources, than by root system size per se.
  • 3Root competition intensity between individual plants generally decreases as resource availability (but not necessarily habitat productivity) increases, but the importance of root competition relative to other factors that structure communities may increase with resource availability.
  • 4Soil organisms play important, and often species-specific, roles in root interactions.
  • 5The findings that some roots can detect other roots, or inert objects, before they are contacted and can distinguish between self and non-self roots create experimental challenges for those attempting to untangle the effects of self/non-self root recognition, self-inhibition and root segregation or proliferation in response to competition. Recent studies suggesting that root competition may represent a ‘tragedy-of-the-commons’ may have failed to account for this complexity.
  • 6Theories about potential effects of root competition on plant diversity (and vice versa) appear to be ahead of the experimental evidence, with only one study documenting different effects of root competition on plant diversity under different levels of resource availability.
  • 7Roots can interact with their biotic and abiotic environments using a large variety of often species-specific mechanisms, far beyond the traditional view that plants interact mainly through resource depletion. Research on root interactions between exotic invasives and native species holds great promise for a better understanding of the way in which root competition may affect community structure and plant diversity, and may create new insights into coevolution of plants, their competitors and the soil community.


Root competition is an almost ubiquitous feature of terrestrial plant communities, with between 70% and 85% of published root competition studies reporting significant results (Wilson 1988; Coomes & Grubb 2000). A century after Fricke's (1904) first publication of a root competition experiment, ecologists no longer ask whether root competition affects plant community structure – and thereby ecosystem functioning – but how important such effects may be relative to other ecological factors that structure plant communities, determine growth and survival of individual plants, and affect plant evolution.

Species composition affects the nature and intensity of root interactions in a plant community, and root interactions can, in turn, affect local species diversity through local competitive exclusion, niche partitioning and facilitation. Root interactions are now at the heart of a debate about the relationships between plant species diversity, ecosystem functions such as primary productivity and community invasibility (Wilson & Tilman 2002; Cahill 2003; Rajaniemi 2003b; Rajaniemi et al. 2003). Changes in species diversity often result in altered ecosystem functions, such as changes in productivity, and carbon, nutrient and hydrological cycles (Chapin et al. 2000; Loreau et al. 2002). Understanding such relationships requires an understanding of plant interactions, of which root interactions are at least a major, and sometimes a dominant, component.

Since the year 2000, five major themes have emerged in research on root competition: (i) effects of soil heterogeneity and its potential for causing size-asymmetric root competition, (ii) effects of resource availability, (iii) the role of soil organisms in root interactions, (iv) the role of signals or toxins in the interaction of roots with other roots and with the soil environment, and (v) relationships between root competition and plant diversity. For summaries of the earlier literature the reader is directed to Caldwell & Richards (1986), Casper & Jackson (1997) and de Kroon et al. (2003) for comprehensive reviews and to more specialist reviews of spatial or temporal root segregation (Fitter 1987; Schenk et al. 1999) and of root competition in forests and woodlands (Nambiar & Sands 1993; Coomes & Grubb 1998), agroforestry systems (Schroth 1999), and arid and semi-arid environments (Fowler 1986). The technical challenges of experimentally separating root and shoot competition have been discussed by Wilson (1988), McPhee & Aarssen (2001) and Cahill (2002b). The effects of soil organisms on root interactions were reviewed in a recent paper by Wolfe & Klironomos (2005).

What is root competition?

Competition is a familiar concept in biology, but there is no universal agreement on its definition (McIntosh 1992). The one used here is that root competition is a reduction in the availability of a soil resource to roots that is caused by other roots (adapted from a general definition of competition by Ricklefs & Miller 1999). Soil resources that roots compete for include nutrients, water and space (Casper & Jackson 1997), and their availabilities may be reduced due to depletion or due to a variety of direct root interactions (see below).

root competition: effect and response

Goldberg (1990) defined plant competition as a process that includes the effect of one plant on a resource and the response of another plant to the change in resource availability. This inclusion of the response conflicts with the way most ecologists use the term ‘competition’. A search in the JSTOR database ( in February 2005 found 225 ecological papers that used some permutation of the phrase ‘response to competition’, implying that the two aspects are separate. However, root competition research can hardly be conducted without studying the responses, which can range from physiological, developmental, growth or behavioural responses of individual plants and roots to community- or ecosystem-level responses, depending on the organizational scale of the study (Table 1).

Table 1.  Variables that may be used as measures of a reduction in the availability of a soil resource to roots that is caused by other roots, i.e. as measures of root competition, and parameters that are potential measures of responses to root competition at different organizational levels. Measurements of root competition and responses at higher organizational levels could also include measurements at lower levels
Organizational scaleDisciplineMeasures of root competitionMeasures of responses to root competition
Individual rootsEcophysiologyUptake and soil concentrations of nutrients and water in rhizospheres of individual roots; root exudates; local soil modificationsRoot anatomy and morphology, elongation rate, curvature, branching, life span
Whole plantEcophysiologyUptake and soil concentrations of nutrients and water in rhizospheres of whole plants; root exudates; local soil modificationsCarbon and nutrient allocation, esp. root/shoot; nutrient acquisition; water use; plant growth and development, plant architecture, incl. spatial distribution of roots
Evolutionary ecologyUptake and soil concentrations of nutrients and water in rhizospheres of whole plants; root exudates; local soil modificationsAll components of plant fitness: fecundity, growth, and survival
PopulationEvolutionary ecology or population ecologyUptake and soil concentrations of nutrients and water in soil occupied by population; root exudates and soil modificationsPopulation growth, demography, including recruitment, age-structure, size structure, survival
Plant communityCommunity ecologyNutrient cycles and soil water balance in the community; root exudates and soil modificationsPlant species diversity, spatial community structure (above ground and below ground)
Whole communityCommunity ecologyNutrient cycles and soil water balance in the community; root exudates and soil modificationsResponses of animals, fungi, protists, bacteria, especially in the rhizosphere; species diversity of all organisms, spatial community structure
EcosystemEcosystem ecologyEcosystem nutrient cycles and soil water balance; root exudates and soil modificationsNet primary productivity; nutrient cycling; hydrology; secondary productivity

types of root competition: contest and scramble

Root competition can affect the availability of a resource to plants either by resource depletion (scramble or exploitation competition) or by mechanisms that inhibit access of other roots to the resource (contest or interference competition) (Fig. 1; Schenk et al. 1999; de Kroon et al. 2003). Traditionally, plant ecologists have used the term allelopathy for contest competition, in some cases considering it to be a distinct process (e.g. Goldberg 1990; Harper 1961, 1977; Muller 1969), but this dichotomy has become obsolete. Plants are now known to be capable of detecting and interacting with neighbours in multiple ways, through effects on resources, exchange of various kinds of signals and allelochemical interactions (Callaway 2002). All of these processes have analogues in animal ecology, where they are considered to be part of competition (Schoener 1983; Ricklefs & Miller 1999).

Figure 1.

The different modes (size-symmetric and size-asymmetric) and types (scramble or contest) of competition that are theoretically possible at different hierarchical levels (intraplant, intraspecific, interspecific). Plant illustrations by Ivy Livingstone are reproduced with permission from the BIODIDAC image collection (

negative and positive root interactions

As well as negative (competitive) interactions, roots may have positive effects on other roots (Callaway 1995; Hauggaard-Nielsen & Jensen 2005). Such root facilitation may be defined as an increase in the availability of a soil resource to roots that is caused by other roots. Potentially, roots could simultaneously decrease availability of one soil resource (e.g. phosphate) and increase availability of another (e.g. water through hydraulic redistribution) (Holzapfel & Mahall 1999).

modes of root competition: size-symmetry and size-asymmetry

If size confers a competitive advantage over a smaller individual, then the larger organism may acquire disproportionately more resources than expected from the size difference between competitors. Such size-asymmetric competition commonly occurs in animals (Schoener 1975) and in shoot competition between plants (Weiner 1985), but it is not clear whether it occurs in root competition (Schwinning & Weiner 1998). Figure 2 illustrates the concept of size-asymmetry for two hypothetical root competitors in comparison with other modes of root competition. For division of resource uptake by competing root systems in a shared soil volume to be proportionate to the size of the root systems, they must have similar resource acquisition efficiencies (RAE, i.e. the amount of soil resource acquisition/costs of soil resource acquisition) (Fig. 2a). Division of resource uptake between competing root systems will be disproportionate to root system sizes if they differ in RAE (Fig. 2b; Yanai et al. 1995) or if contest competition, such as chemical inhibition of a competitor's roots, takes place (Fig. 2d), as well as if root competition is size-asymmetric (Fig. 2c). Note that, contrary to some statements in the literature (e.g. Crawley 1997; Bauer et al. 2004), size-asymmetric competition is not synonymous with contest competition: both scramble and contest competition can, theoretically, be either size-asymmetric or size-symmetric (Fig. 1).

Figure 2.

Soil resource acquisition for two hypothetical root competitors, X and Y, that share a soil volume without additional competitors. (a) Size-symmetric competition with similar resource acquisition efficiencies (RAE, i.e. the amount of soil resource acquisition/costs of soil resource acquisition) for X and Y. (b) Size-symmetric competition; plant Y has higher RAE than plant X. (c) Size-asymmetric competition. (d) Contest competition, such as chemical inhibition of a competitor's roots. See text for further explanations.

Size-asymmetric competition occurs if larger size confers a better chance of a plant acquiring resources that are directional or not distributed homogeneously in space (Schwinning & Weiner 1998). For example, leaves at the top of the canopy can pre-empt the use of light energy by leaves lower in the canopy and it is the amount and spatial distribution of resource-acquiring organs, relative to the directionality of resources, that confers the competitive advantage, rather than increased plant size per se. Plants of the same size (e.g. height, volume, mass) can have very different distributions of resource-acquiring organs, which means that the concept of size-asymmetry may be problematic, especially when applied to interspecific competition.

methods in root competition research

Given our definition of root competition, its intensity is best measured by comparing soil resource availability in the presence and absence of competing roots. Acquisition rates of specific resources can be measured (e.g. Yuan et al. 2004), as can uptake of tracers, such as nutrient analogues (Casper et al. 2000, 2003), stable isotopes (e.g. Berntson & Wayne 2000) or radioactive tracers (e.g. Caldwell et al. 1985). The use of genetically engineered bacteria as biosensors of specific compounds or water potential around roots holds particular promise for research on interactions between individual roots (Z. Cardon, personal communication). Plant responses to root competition (including changes in yield, relative growth rate, fecundity, survival and net primary productivity) are much more easily and commonly measured than resource availability itself (Goldberg 1990). The choice of response parameters to be measured depends on the question asked and the hierarchical level (individual, community or ecosystem) of the study (Table 1; McPhee & Aarssen 2001; Cahill 2002b; Zobel & Zobel 2002). Interactions between root and shoot competition occur frequently, and, unless otherwise noted, the necessary experimental separation of above- and below-ground effects (McPhee & Aarssen 2001; Cahill 2002a) has been carried out in the root competition studies reviewed here.

Effects of soil heterogeneity on root competition

The effects of soil nutrient heterogeneity on performance of individual plants, populations and species mixtures have recently been reviewed by Casper et al. (2000), Hutchings et al. (2003) and Hodge (2004). According to a predictive framework developed by Hutchings et al. (2003), the performance of individual plants, plant populations and communities on heterogeneous soil will depend on the sizes of both nutrient-rich patches and root systems, the distances between plants and nutrient patches, and the contrast in fertility between patches and other parts of the soil. Strong competition is predicted to occur within nutrient-rich patches and at borders of patches. More recent studies have found that nutrient-poor patches in heterogeneous soil, which contained lower root mass, allowed plants to survive better than in nutrient-rich patches, probably by providing some release from root competition (Day et al. 2003a), and that experimental communities grown on heterogeneous soil had more root biomass than those on homogeneous soil (Wijesinghe et al. 2005). Soil heterogeneity in the latter study had a variety of effects on biomass, cover or population size of different species, but no effects on species richness or diversity of the communities generated.

Only two of the numerous studies on effects of nutrient heterogeneity on plant competition (Blair 2001; von Wettberg & Weiner 2003) appear to have experimentally isolated root competition from shoot competition, and these found no effects of nutrient heterogeneity on root competition intensity. By contrast, two grass species, Briza media and Festuca ovina, grown in pots with homogeneous or heterogeneous nutrient supply, were found to have higher interspecific competition intensity under heterogeneous conditions, possibly due to the selective placement and high concentration of roots in nutrient patches (Day et al. 2003b). Similar effects of nutrient heterogeneity on competition intensity were found for Molinia caerulea in competition with Carex hartmanii (Janeček et al. 2004). These and other reports (see Hutchings et al. 2003) of effects of nutrient heterogeneity on overall competition may reflect effects of soil heterogeneity on shoot growth and interactions between root and shoot competition.

Soil water is a resource with a high degree of spatial and temporal heterogeneity, especially in water-limited environments. In arid and semi-arid climates, shallow water infiltration is the most common and therefore the most reliable, but most short-lived, source of water. Medium infiltration is more rare and deep infiltration most rare, most long-lived, but least reliable (Loik et al. 2004). Groundwater is a reliable but spatially restricted source that is available only to a small number of highly specialized, phreatophytic species. Plants in dry lands may tap into one or all of these pools and differ widely in their use of water in wet and dry seasons (Chesson et al. 2004), depending on the shape of their root systems, their phenologies and their growth forms (Fig. 3; Schenk & Jackson 2002). Although intraspecific root competition for water can be intense (e.g. Schenk & Mahall 2002), differences between species in root system sizes and shapes and in root phenology can potentially result in opportunities for hydrological niche differentiation between coexisting species (Walter 1939; Silvertown et al. 1999; Chesson et al. 2004; Ogle & Reynolds 2004).

Figure 3.

Median root system sizes and canopy sizes of different plant growth forms from global arid environments. The data were compiled from the literature by Schenk & Jackson (2002). The graph illustrates median and maximum rooting depths (n = 750), median and maximum lateral root spreads (n = 526), median depth of maximum lateral root spreads (n = 239), median canopy heights (n = 362) and median canopy widths (n = 336).

size-asymmetry in root competition

Most experimental evidence suggests that resource acquisition by root systems of competing plants tends to be roughly proportional to the differences in their sizes (Weiner 1986; Schwinning & Weiner 1998; Berntson & Wayne 2000; Cahill & Casper 2000). It has been hypothesized, however, that in heterogeneous soils a root system's chance of encountering and using patches of soil resources may increase non-linearly with its size, potentially causing size-asymmetric root competition (Schwinning & Weiner 1998). Some evidence to support this prediction was found for intraspecific root competition in the grass Bromus inermis grown in pots filled with unhomogenized (and therefore presumably heterogeneous) field soil (Rajaniemi 2003a) and for Trifolium subterraneum grown in trays with a patchy distribution of phosphorus (Facelli & Facelli 2002). Fransen et al. (2001) observed larger, but temporary, above-ground shoot size-inequalities in two grass species when growing on heterogeneous soil, which they interpreted to be indirect evidence for size-asymmetric root competition. In a dense monoculture of Helianthus annuus grown in natural field soil, nitrogen uptake rates increased exponentially with above-ground plant sizes (Yuan et al. 2004), although this may have been caused by size-dependent changes in root/shoot allocation. Other experiments produced no evidence for size-asymmetric root competition in heterogeneous soil (Blair 2001; Rajaniemi 2003a; von Wettberg & Weiner 2003). Root competition could indirectly contribute to size-asymmetric shoot competition if a larger root system enabled a plant to grow taller than its competitors (Cahill 1999).

The nature of soil heterogeneity (contrast and spatial scale of patches in relation to the scale of the root system or size of depletion zones) and the overall availability of resources, as well as the number, sizes, kinds and spatial arrangement of competitors, will determine whether the costs of producing a larger root system are offset by a sufficiently higher resource acquisition (Hutchings et al. 2003) and thus affect the size-asymmetry of root competition. A recent modelling study found that size-asymmetric root competition may be difficult to detect in dense, productive communities where roots of an individual plant interact with those of dozens of neighbours of different sizes (Bauer et al. 2004).

Root competition for soil resources with a directional component provides an intriguing analogue to shoot competition for directional light. Soil resources typically have a strong vertical dimension, with a directional component provided by soil water infiltration (Schenk 2005). Roots that are closer to the surface generally have a competitive advantage over deeper roots, because, on average, they have access to larger amounts of water and nutrients (van Wijk & Bouten 2001). This is presumably why, globally, the upper 0.2 m of the soil, including litter layers, contains almost 60% of all roots (Schenk 2005). Vertical root distributions often change in response to root competition, and several studies have found that superior competitors have higher proportions of their roots in the uppermost soil layers (Bunce et al. 1977; D’Antonio & Mahall 1991; Leuschner et al. 2001; Genney et al. 2002). A major difference between root- and shoot-competition for directional resources is that plants of any size have potential access to the upper soil layers, while reaching the top of the canopy can be costly or impossible for many plants. Any type of competition for vertically distributed, directional resources is likely to shape the vertical distribution of resource-acquiring organs (Schenk 2005), which may or may not be related to attributes of ‘plant size’.

Effects of soil resource availability on root competition

A meta-analysis by Goldberg et al. (1999) showed that effects of both negative and positive plant interactions on plant growth tend to be stronger in resource-poor than in resource-rich environments. Negative effects of competition on plant survival tend to be strongest under resource-poor conditions (Goldberg et al. 1999), where relatively small additions or removals of resources are likely to have disproportionally large effects on growth and survival of other plants. The only two studies that have directly measured root competition intensity in environments that differ in resource availability both concluded that root competition was more intense under resource-poor conditions (Cahill 1999; Pugnaire & Luque 2001), and the same conclusion has been derived from indirect measures (Wilson & Tilman 1993; Peltzer et al. 1998; Rebele 2000; Morris 2003). However, others have found root competition to be equally intense over a range of resource availabilities (Belcher et al. 1995; Twolan-Strutt & Keddy 1996).

Root competition in forests tends to be more intense when conditions are nutrient-poor or water-limited, and for understorey plants growing in deep shade vs. those in forest gaps (Coomes & Grubb 2000). The increase in root competition as crucial resources become increasingly limiting does not, however, necessarily mean that it increases with decreasing habitat productivity, as demonstrated by intense root competition among understorey plants in productive forests. Root competition for water in dry environments may be more important for structuring plant communities than root competition for nutrients on infertile but moist soils, because a lack of water tends to kill many species whereas a lack of nutrients in natural environments merely tends to stunt plant growth (Coomes & Grubb 2000). Severe infertility is sometimes associated with soil toxicity (such as aluminium toxicity or salinization), which is likely to affect root competition between species that differ in their toxicity-tolerance.

The general negative relationship between root competition intensity and resource availability seems to be fairly well established, although with the expected variance between studies and systems. Much less information is available regarding the importance of competition (sensuWelden & Slauson 1986) relative to other factors that affect plant growth, reproduction and survival along gradients of resource availability. In a re-analysis of data from Reader et al. (1994) and Pugnaire & Luque (2001), the importance of total (shoot + root) competition was found to increase with standing crop and thereby with resource availability (Brooker et al. 2005). If the importance of root competition also tended to increase with resource availability, negative effects of root competition on plant growth, reproduction and survival, and thereby on plant diversity, could potentially be more important under resource-rich than under resource-poor conditions (Fig. 4). Such a pattern was found for experimental old-field communities, where negative effects of root competition on diversity were strong, and increased when resources were added through fertilization (Rajaniemi et al. 2003).

Figure 4.

Hypothesized effects of root competition on plant performance along a resource gradient. The figure illustrates the intensity of root competition (Cint) and the importance of root competition relative to other performance-limiting factors (Cimp) (Brooker et al. 2005). P(T + N) is the performance of a target plant in the presence of root competition from neighbours, P(T − N) is the performance of a target plant in the absence of root competition from neighbours, and max[P(T − N)] is the maximum value of P(T − N) along the gradient, here set to 100%. Rtot is the total reduction in plant performance, and Rrootcomp is the reduction due to root competition. In this example, Cint is −0.4 and −0.1 and Cimp is −0.29 and −1.00 at the low and high end of the resource gradient, respectively, so that root competition is more intense at low resource availability but more important for limiting plant performance at high resource availability.

Effects of soil organisms on root competition

Mycorrhizal associations affect availability of resources such as phosphates, and can therefore affect competitive interactions (Zobel et al. 1997). The presence of arbuscular mycorrhizae increased intraspecific size inequalities between competing plants of Hypericum perforatum (Moora & Zobel 1998) and Trifolium subterraneum (Facelli & Facelli 2002), possibly by enabling larger plants to monopolize patches of limiting soil resources, and thus promoting competitive size-asymmetry. A large increase in plant species diversity following suppression of mycorrhizal symbiosis in a tallgrass prairie community was attributed to different responses of the dominant and subdominant species to infection (Hartnett & Wilson 1999). Arbuscular mycorrhizae strongly favoured invasive non-native species of Centaurea in competition with North American native perennial grasses (Marler et al. 1999; Callaway et al. 2001, 2003). Centaurea maculosa was able to gain more phosphorus from its mycorrhizal symbionts when grown in competition with Festuca idahoensis, although the reasons for this effect remain unknown (Zabinski et al. 2002). Mycorrhizae also enabled Prunus caroliniana seedlings to compete more successfully with an exotic invasive shrub, Ardisa crenata (Bray et al. 2003).

Interactions between roots and pathogenic soil fungi can also affect plant interactions and result in apparent competition (Connell 1990). Soil fungi associated with the annual legume Chamaecrista fasciculata were detrimental to the perennial grass Andropogon gerardii, but did not affect Chamaecrista (Holah & Alexander 1999). Pathogenic soil fungi associated with adult black cherry trees negatively affected conspecific seedlings growing within their root zone (Packer & Clay 2000), but only in their native North America (the fungi may not occur in Europe, where black cherry is an invasive exotic; Reinhart et al. 2003). Species-specific pathogenic soil fungi can accumulate in the rhizosphere and cause species replacement, thereby contributing to successional change and increasing plant diversity (Bever 1994; Mills & Bever 1998).

Root exudates can influence the composition of the bacterial flora in the rhizosphere of plant species (Reynolds et al. 2003; Sturz & Christie 2003). Soil bacteria can potentially increase or decrease nutrient availability in the rhizosphere, which may in turn affect nutrient availability to competing roots. Species-specific associations between soil fungi or microbes and plant roots may be products of co-evolution and may play integral parts in root interactions between species (Fitter 2005; Wolfe & Klironomos 2005).

Root interactions involving signals or toxins

There is mounting evidence that roots can detect (and react to) not only the abundance of soil resources, but also the presence (and even the identity) of other roots in the soil, as well as the presence of inert objects, such as the wall of a soil container, before they even make contact. Detection mechanisms appear to involve non-toxic signals and/or toxic allelochemicals.

interactions of roots with inert objects

Soil is a very heterogeneous medium that contains solids of various sizes, air spaces and soil solution. Resource-acquiring organs in such a medium may have an advantage if they can detect the presence of large solid objects (or air spaces) that are unlikely to yield resources, and react by reducing or redirecting growth towards them. Recent research has shown that pea roots are able to detect barriers in their growth medium even before physically contacting them, possibly by using a kind of ‘chemical radar’ that causes self-inhibition of root growth when approaching mechanical barriers (Falik et al. 2005). Similar mechanisms may be responsible in part for strong effects of restricted soil volume on plant growth and development, long before there are any indications of pot binding (McConnaughay & Bazzaz 1991; NeSmith & Duval 1998).

signals and toxins in intraplant root competition: self-inhibition

Competition between roots of the same root system will reduce RAE and could reduce a plant's fitness. Uptake of immobile resources, such as phosphate, is likely to be more affected by intraplant root overlap than uptake of more mobile resources, such as nitrate or water, which depends less on exactly where the roots are placed. These postulates were tested in a study that compared wild-type Arabidopsis thaliana with a mutant that had a reduced number of lateral roots. Phosphate, but not nitrate, affected the fitness of the mutant relative to the wild-type (Fitter et al. 2002). A detailed modelling study of root system architecture in bean plants also came to the conclusion that intraplant root competition can greatly affect the acquisition of phosphate (Ho et al. 2004).

In theory, a plant may increase its relative fitness by growing roots into soil shared with a competitor, allowing it to acquire resources from that space and also, potentially, to decrease the competitor's fitness by reducing its resource supply (Robinson et al. 1999). A plant may therefore benefit from being able to detect the difference between ‘self’ and ‘non-self’ roots, and to inhibit growth of self roots once they start to intermingle. Studies on a variety of herbaceous and woody plant species (Mahall & Callaway 1996; de Kroon et al. 2003; Falik et al. 2003; Holzapfel & Alpert 2003; Gruntman & Novoplansky 2004) have shown that roots behave differently when encountering ‘self’ vs. ‘non-self’ roots, rather than simply responding to resource depletion zones. The nature of the signals remains unknown, but the lack of harm suggests that they are not toxins (Mahall & Callaway 1992; Schenk et al. 1999; Falik et al. 2003; Holzapfel & Alpert 2003). They may be of an electrical nature (Schenk et al. 1999; Falik et al. 2003; Gruntman & Novoplansky 2004). Other plant species may limit intraplant root overlap by exuding allelochemicals that inhibit root growth, as reported for the desert shrub Larrea tridentata (Mahall & Callaway 1992).

signals and toxins in intraspecific root competition

Conspecific plants are, by definition, similar in their demands and abilities with regard to resource acquisition. Despite this, the available evidence suggests that intraspecific competition for resources is not necessarily more intense than between species (Goldberg & Barton 1992). Recent research has found evidence that intraspecific competition may be reduced, particularly in clonal species, by reducing overlap between the roots of conspecific neighbours (Mahall & Callaway 1991, 1992, 1996; Schenk et al. 1999; de Kroon et al. 2003). Such mechanisms may increase genet fitness by reducing intraclonal competition between genetically identical individuals or even between physically connected but physiologically independent plant modules (Schenk 1999). Another mechanism that results in reduced root competition between genetically similar plants was recently reported for Centaurea maculosa, in which root exudation of (+/–)-catechin reduces conspecific seedling establishment near adult plants (Perry et al. 2005).

signals and toxins in interspecific root interactions

Toxic, and as yet unidentified, allelochemicals exuded by roots of the desert shrub species Larrea tridentata have been shown to inhibit growth of its own roots as well as roots of a common neighbour, the shrub Ambrosia dumosa (Mahall & Callaway 1992). The most complete study of interspecific root interactions involving toxic allelochemicals to date is that of Bais et al. (2003), who measured exudation rates of the phytotoxin (–)-catechin from roots of the invasive exotic species Centaurea maculosa. The data showed inhibition of native species’ growth and germination in field soils at natural concentrations of the toxin, and demonstrated the toxin's effects on a target plant species at physiological and molecular levels. There are indications that native species may evolve defences against the toxin through natural selection (Callaway et al. 2005). Allelochemicals may play an important role in allowing the spread of invasive species that possess ‘novel weapons’ that the native plants have not previously encountered, but they may be less important for structuring communities of co-evolved species (Callaway & Aschehoug 2000; Hierro et al. 2005).

signals, toxins and self-inhibition in root interactions: new challenges for experimental design

Differential responses to inert objects, roots from the same physiological individual, roots from other conspecific individuals and roots from other species create a new challenge for experimental design, especially for root competition studies conducted in pots.

Many root competition studies use variations on the ‘divided pots technique’ (McPhee & Aarssen 2001), in which barriers divide the soil space in ‘no root competition’ treatments. A frequently used variant of this technique (Donald 1958) keeps the absolute soil volume constant in all treatments, but varies plant density and thereby the soil volume available per plant (e.g. Aspinall 1960; Lambert 1967; Kitamura et al. 1981; Haugland & Froud-Williams 1999). Gersani et al. (2001) introduced a new ‘split root’ design, in which plants were grown with their roots divided between two pots in the presence of an equally divided competitor, or in only one pot without a competitor (treatments 2 and 3 in Fig. 5). This arrangement keeps plant density – i.e. number of plants per soil volume – constant, ensuring that the same amount of soil resources is theoretically available per plant. Previous studies achieved the same goal by varying the sizes of soil compartments between ‘root competition’ and ‘no root competition’ treatments (e.g. Litav & Wolovitch 1971; Litav & Isti 1974; Fitter 1977; Arnone & Kestenholz 1997; Xiao et al. 2004). In these studies, as well as those of Gersani et al. (2001), Maina et al. (2002), Falik et al. (2003) and O’Brien et al. (2005), treatments differed not only in the presence or absence of root competition, but also in the absolute sizes (and sometimes shapes) of soil volumes available for root growth of each plant. Such a design fails to account for the fact that the absolute size and shape of the soil volume can strongly affect plant growth and development (Hanson et al. 1987; McConnaughay & Bazzaz 1991; van Iersel 1997; NeSmith & Duval 1998) and that species respond differently to competition in different sized pots (Taylor & Aarssen 1989; Gurevitch et al. 1990).

Figure 5.

Experimental design of the root competition studies of Maina et al. (2002) and the author (unpublished data; see Figure 6). Split root systems were created in legume seedlings as described by Gersani et al. (1998, 2001). (1) Two plants rooted together in one pot, (2) two plants rooted together in two adjacent pots, (3) two plants separately rooted in two adjacent pots and (4) two plants with root systems separated and each growing in two adjacent pots.

In fact, if soil volume is taken into account, data from Maina et al. (2002) and Gersani et al. (2001) and my unpublished data show that root mass of legume plants increases in response to an increase in absolute soil volume, but is not affected by the presence of a competitor (Fig. 6). This calls into question the conclusion reached by Gersani et al. (2001), Maina et al. (2002) and O’Brien et al. (2005) that the increase in root mass was a response to detecting that resources were shared with a competitor, i.e. a ‘tragedy of the commons’ (Hardin 1968). By contrast, numerous previous studies had found a reduction in root mass in response to competition (e.g. Lambert 1967; Fitter 1977; Bozsa & Oliver 1990; Haugland & Froud-Williams 1999; Semere & Froud-Williams 2001). The patterns illustrated in Fig. 6 clearly suggest that root growth was responding not to the presence of a competitor but to the larger soil volume, possibly as a result of reduced self-inhibition between roots (Falik et al. 2003; Gruntman & Novoplansky 2004) and of the increased distances over which roots could grow (Falik et al. 2005). Nevertheless, self-recognition and self-inhibition between roots clearly can affect root competition with other plants, and these findings open up new avenues of research into root behaviour.

Figure 6.

Root dry mass for legume plants with split root systems grown in pots with or without intraspecific root competition (treatments illustrated in Fig. 5). Square symbols, with intraspecific root competition; round symbols, without intraspecific root competition. The study by Gersani et al. (2001) lacked treatments 1 and 4 from Fig. 5, but used two different types of containers. My study was conducted as described by Maina et al. (2002). It had five replicates in each treatment. Above-ground parts of each pair of plants grew in close proximity, but did not shade each other. The plants were grown for 6 weeks in the glasshouse of the California State University Fullerton and were watered twice-weekly with MaxiGrow™ nutrient solution (0.25 g L−1; General Hydroponics, Sebastopol, CA, USA).

Root competition and plant diversity

Much of the literature on plant diversity is based on the theory that plants can coexist because of niche-partitioning or complementarity (Hooper 1998; Loreau & Hector 2001), whereby traits such as differences in rooting depth, root system architecture or root foraging behaviour allow species to avoid competitive exclusion by exploiting different fundamental niches (Cody 1986; Silvertown 2004; Stubbs & Wilson 2004). However, the fundamental niches of many plant species overlap greatly – and often completely – because plants have far fewer options for making a living than heterotrophic organisms. Interactions between species cause a fundamental niche to be partitioned into realized niches (Hutchinson 1957), displacing many species to suboptimal parts of their fundamental niches (Ellenberg 1953; Austin & Austin 1980; Austin 1999).

Until recently, plant ecologists generally assumed that plants mainly interact indirectly through resource depletion, but we now also recognize positive effects (Callaway 1995; Hauggaard-Nielsen & Jensen 2005), various kinds of signals and recognition systems (Callaway 2002; Bais et al. 2004; Gruntman & Novoplansky 2004; Falik et al. 2005), toxins (Bais et al. 2003) and indirect root interactions that involve soil organisms (Callaway et al. 2004; Fitter 2005; Wolfe & Klironomos 2005). Biotic interactions may therefore play an important role in shaping realized niches of plant species, implying that abiotic effects alone cannot explain plant diversity. Attempts to do so (e.g. Stubbs & Wilson 2004) potentially miss the most important factor shaping species coexistence, namely species interactions (Schenk 1996), of which root interactions are a major – often a dominant – component.

If plant interactions truly are important for structuring plant communities, then root competition must play a major part. The question is: How important is root competition for determining plant diversity and how does it vary with resource availability (and thereby often with productivity)? The only study to address this question directly (Rajaniemi et al. 2003) found that negative effects of root competition on plant diversity increased with increasing nutrient availability. A potential problem with research on the effects of plant interactions on community structure is that current interactions between coexisting species may not be sufficient to determine the role of interactions earlier in community development: species excluded by past interactions are no longer there to tell their tale. Possible solutions to the problem of which other species to consider are to use a phytometer approach, growing the same species in competition with species that naturally occur in a range of different environments (e.g. Reader et al. 1994), or to create artificial communities of the same group of species under a range of environmental conditions (e.g. Hector et al. 1999; Fridley 2003). To date, neither of these approaches has been used specifically to study the intensity and importance of root competition along gradients of resources or environmental stress.

The effects of root competition on plant diversity can also be investigated by experimentally reducing root competition in a natural environment (e.g. by trenching), as in many studies of forest understorey communities (beginning with Fricke 1904 and reviewed by Coomes & Grubb 2000). Microhabitats with reduced root competition may also exist in nature. Rocky, shallow soils over fragmented bedrock may provide plants with pockets of soil that are largely free of root competition, resulting in higher plant diversity than on deeper soils (Hölscher et al. 2002). Canopy gaps can provide niches for poor root competitors (Cahill & Casper 2003) and nutrient-poor patches of soil may also be places that allow poor competitors to survive (Day et al. 2003a).

Exotic, invasive species that do not share a co-evolutionary history with the native vegetation offer unique opportunities to study the role of plant interactions in shaping community structure (Callaway & Ridenour 2004; Vila & Weiner 2004; Wolfe & Klironomos 2005). Negative interactions between the roots of exotic and native species are often more intense than those between species that have coexisted for a longer period of time, sometimes involving allelochemicals (‘novel weapons’) against which native plants have not evolved defences (Callaway & Aschehoug 2000; Ridenour & Callaway 2001; Bais et al. 2003; Hierro & Callaway 2003; Callaway et al. 2005; Hierro et al. 2005), and species-specific interactions with soil organisms can also play an important role in below-ground interactions between invasive plants and their competitors (Callaway et al. 2001, 2003; Zabinski et al. 2002; Ridenour & Callaway 2003).

Root competition can potentially exclude invaders, but is this more likely in diverse communities, as postulated by Levine & D’Antonio (1999)? The only study designed so far to answer this question found no effect of diversity of competitors on the intensity of root competition in old field vegetation (Cahill 2003). Rajaniemi (2003b) proposed that size-asymmetric root competition may limit plant diversity under resource-poor conditions, based on the postulate that resource-poor soils tend to be more heterogeneous. However, there is only limited experimental support for size-asymmetric root competition, and it is by no means clear what types of soil heterogeneity it may require. Wijesinghe et al. (2005) found no effects of soil heterogeneity on species richness and diversity in experimental communities. More studies are required to determine whether such effects may be important in nature.

Synthesis and future research needs

Research on root competition has come a long way since Fricke (1904) first published results from a soil trenching experiment to determine whether forest understorey plants were limited by soil resources or by light. Interestingly, the general question asked by plant ecologists today about the role of root competition on diversity and structure of plant communities is still the same as that asked by Fricke, and we do not appear to be much closer to an answer. Soils are complex media, and root competition is likely to be affected by complex interactions between the spatial heterogeneity of soil resources and the spatial distribution of competing resource-acquiring organs below ground (roots and mycorrhizal hyphae). It seems unlikely that the theory of size-asymmetric root competition on heterogeneous soils (Schwinning & Weiner 1998; Rajaniemi 2003b) will be able to explain much about relationships between plant diversity and resource availability, largely because size is a hard-to-define concept with regard to modular organisms such as plants, and because all natural soils are more or less heterogeneous.

The choice of appropriate model systems for studies of root competition is an issue that deserves far more attention. Crop and agroforestry plant models are convenient, often well studied and easy to handle, but selection for varieties that are not strong root competitors when grown at high density with conspecific neighbours may limit their use for predicting responses of wild plants (Denison et al. 2003).

Plant diversity in a given environment cannot be explained entirely by characterizing resource availability and the demands and resource-acquisition capabilities of plant species, or by studying plant responses to abiotic stress (Ellenberg 1953). To understand better the mechanisms that shape plant diversity, we have to move beyond the idea that plants interact mainly through resource depletion. New findings about signalling mechanisms and chemical interactions between roots, and about species-specific interactions of roots with soil organisms, suggest that we may need to understand much more about the mechanisms by which roots interact with their environment. More specifically, we need to understand how a root apex ‘makes decisions’ about when and where to grow and when to cease growth, and what information about the abiotic and biotic environment can be detected and processed by a root (Baluška et al. 2004). Moreover, we will have to consider the possibility of co-evolution between plant competitors (Hierro et al. 2005) and between plants and soil organisms (Callaway et al. 2004). Research on invasive plant species provides an opportunity to study root interactions in species that have not co-evolved with their competitors and with the soil community. For once the needs of ‘applied’ and ‘basic’ research overlap, and the next decade promises to be a fruitful period for research on root interactions.


Many thanks to Claus Holzapfel and J.C. Cahill for comments on earlier drafts of the manuscript and to Lindsay Haddon for editing the final version. The helpful comments of Mike Hutchings and two anonymous referees are greatly appreciated. I wish to acknowledge the contributions of students in the plant ecology class in autumn 2004 at the California State University Fullerton – Daniel Bergin, Susana Espino, Dawn Hendricks, Peter Koenig, Daravone Laliemthavisay, Patricia Lambert, Steven Peterson, Joy Polston Barnes, Dallas Tognotti, Nicole Vearrier and Andrea Vona – who helped to conduct the experiment described in Figs 5 and 6. Support for this research was provided by a grant from the Andrew W. Mellon Foundation.