Cluster roots: some ecological considerations


Keith R. Skene (fax 01382 344275; e-mail


It has been said that the ‘roots of ecology are in the ecology of roots’ (Sen 1980). The study of resource acquisition by roots is a vibrant and challenging field today, and an important meeting point between ecology, physiology and developmental biology. Most of the essential elements for all life forms enter the biosphere through the roots of plants, and so the subject demands careful consideration (Nissen 1991). The determinance of plant community structure also has its roots firmly embedded in the soil. While much emphasis is placed on the ecological significance of mycorrhizas and nitrogen-fixing symbioses, it is surprising that so little attention has been paid to another root adaptation that occurs throughout the world, namely the cluster root. In this short article, it is argued that cluster roots are the third great adaptation for nutrient uptake in plants and, as such, demand greater attention from the ecologist than they presently receive.

The term ‘cluster root’ is used to refer to a region of parent root (usually a primary or secondary lateral root) where many short rootlets are produced in a compact grouping, giving the appearance of a bottle brush (Fig. 1) (Purnell 1960). First noted by Engler (1894), the rootlets develop opposite every protoxylem pole and are determinate, the meristem itself eventually differentiating. Their gross morphology differs between different species, in that they may be simple, complex or compound (Fig. 1; adapted from Skene 1997). Compound cluster roots can form large mats in the upper surface of the soil, and, where they are produced by trees, often extend beyond the canopy.

Figure 1.

Cluster root morphology. (a) Cluster roots (CR) are made up of rootlets (Ro) that, at maturity, are covered in rootlet hairs (RH). Note the ellipsoid shape of the mature roots. (b) The simple cluster root, common in non-Proteaceae and in some members of the Proteaceae. (c) A complex cluster root, with a second cluster root emerging from within the first one (occasionally observed in species with simple cluster roots). (d) The compound cluster roots, common in mat-forming members of the Proteaceae.

Phylogeny and distribution

Keywords: Table 1 lists all species so far known to produce cluster roots. They occur in almost all of the 1800 species of the Proteaceae, with the only exceptions being in the primitive genus Persoonia (Lamont 1982). Originally, they were thought only to occur in this family and were called ‘proteoid roots’ (Purnell 1960), but they are now known to occur in the Betulaceae, Myricaceae, Mimosaceae, Fabaceae, Casuarinaceae, Eleagnaceae and Moraceae. Their geographical distribution includes South and North America, Europe, Australia, Africa, Asia and the Pacific islands. In view of their prevalence in some of the major plant families in both temperate and tropical ecosystems, their ecological role needs to be examined. Furthermore, their prevalence in certain ecosystems suggests an important ecological role. For example, cluster roots are widespread in regions of significant biodiversity such as the Fynbos, South Africa and in south-western Australia. Species with cluster roots are also often pioneers in primary or in secondary succession, such as on Fraser Island, Australia, where the youngest dunes are occupied by Banksia and Casuarina (White 1986). Lupinus albus is used as a pioneer crop in land reclamation (Smart 1990). Ecologically, the majority of actinorhizal plants, many of which form cluster roots, are pioneers on nitrogen-poor, open sites (Baker & Schwintzer 1990). Furthermore, cluster roots occur in economically important species such as Lupinus (cattle feed, green manure, coffee substitute), Grevillea (timber) and Macadamia integrifolia (macadamia nuts). The distribution and function of cluster roots was reviewed Dinkelaker et al. (1995). Although known for some time in subtropical regions, the occurrence of cluster roots in temperate ecosystems has only recently come to light, with these structures now found in Myrica gale, Myrica pensylvanica, Comptonia peregrina, Alnus glutinosa, Alnus rubra, Alnus incana and Hippophae rhamnoides (see Table 1 for references). While cluster roots are known to be of particular importance in low-phosphate subtropical ecosystems, species producing cluster roots in temperate ecosystems appear to fill similar niches. Thus Myrica gale occurs in waterlogged soils and Hippophae rhamnoides in sand dune succession, and they may play similar roles to Viminaria juncea, and Casuarina and Banksia spp., respectively, under tropical conditions. However, fieldwork regarding the occurrence of cluster roots in temperate regions such as northern Europe is lacking. It is remarkable that such an important structure has been ignored, but exciting also, in that a whole new field of study lies ahead. Urgent consideration of the role of cluster roots in the ecology of temperate ecosystems is required.

Table 1.  Occurrence of actinorhizal/rhizobial nodulation and mycorrhiza in species with cluster roots
  1. 1, Lamont (1982); 2, Louis et al. (1991); 3, Louis et al. (1990); 4, Hurd & Schwintzer (1997); 5, Hurd & Schwintzer (1996); 6, Reddell et al. (1986); 7, Khan (1993); 8, Bowen (1981); 9,Gardner et al. (1982); 10, Racette et al. (1990); 11, Lamont (1972); 12, Trinick (1977); 13, Sward (1978); 14, Sprent (1995); 15, Clements et al. (1993); 16, Rosenfield et al. (1991); 17, Berliner & Torrey (1989); 18, Molina (1981); 19, Reddell et al. (1997); 20, Brundrett & Abbott (1991); 21 Vierheilig et al. (1994); 22, R. Parsons & K. Skene, unpublished data; 23, Gardner et al. (1984).

All Proteaceae (exception: genus Persoonia)1NoNoNo1
Comptonia peregrina4YesNoRare4
Myrica cerifera3YesNoUnclear3
Myrica gale2YesNoUnclear17
Myrica pensylvanica4YesNoRare4
Alnus glutinosa5YesNoEntire genus very
Alnus incana5YesNospecialized in terms
Alnus rubra5YesNoof symbiont18
Allocasuarina campestris6YesNoIn this family,
Casuarina cunninghamiana7YesNocluster roots
Casuarina equisetifolia8YesNoform in
Casuarina littoralis9YesNoconditions
Casuarina obesa6YesNowhere mycorrhizas
Gymnostoma papuanum10YesNodo not form easily19
Hippophae rhamnoides22YesNoYes23
Acacia mucronata13NoYesUnknown
Daviesia cordata14NoYesNone20
Daviesia decurrens14NoYesNone20
Lupinus albus9Lupinus atlanticus15Lupinus cosentinii12 Lupinus digitatus15Lupinus micranthus15Lupinus palaestinus15Lupinus pilosis15Lupinus princei15No No No No No No No NoYes J Yes G Yes G Yes f Yes F Yes G Yes G Yes j Üra5ÝThe genus Lupinus is thought to be a non-host,21 although there are some reports of weak infection12
Kennedia spp.12NoYesUnknown
Viminaria juncea11NoYesAM/no ecto20
Ficus benjamina16NoNoUnknown

Mycorrhizal status and nodulation in species with cluster roots

Function of cluster roots

Low phosphate levels have been shown to encourage cluster root formation (Walker & Pate 1986; Skene et al. 1996). These roots are thought to improve phosphate absorption in several ways. First, the production of reducing agents, hydrogen ions, chelating agents and phosphatases as root exudates are believed to increase phosphate mobility (Gardner et al. 1982; Grierson & Attiwell 1989; Reddell et al. 1997). Secondly, cluster roots show increased absorption rates for P compared to ordinary roots on a dry weight basis and per unit area of root (Malajczuk & Bowen 1974; Vorster & Jooste 1986; Horst & Waschkies 1987). In the Proteaceae, a non-mycorrhizal family dominant in regions with nutrient-impoverished soils (Johnson & Briggs 1975; Tester et al. 1987), such an adaptation should improve nutrient acquisition. The lack of arbuscular mycorrhizas in plants with cluster roots may indicate that their roles are similar (Lamont 1982). Cluster roots have also been implicated in nitrogen uptake, with higher concentrations of amino acids in xylem sap of cluster roots than of other roots (Pate & Jeschke 1993). N-15-labelled organic sources have been shown to be taken up by Hakea (Turnbull et al. 1996).

Exudation in cluster roots represents a physiological clustering, akin to the morphological clustering observed in their structure. It occurs as an ‘exudative burst’, a remarkable phenomenon wherein the rootlets exude little material until fully grown, but then, over 2–3 days, exude large amounts of citrate and malate (Dinkelaker et al. 1995; Neumann et al. 1995; Skene et al. 1996). Following this, the level of exudation drops back to almost zero. The mechanics and ecological significance of this event are still under investigation, but it may be a ‘light brigade’ adaptation, wherein one large exudative burst might prevent bacteria from metabolizing the exudate completely before it can enhance uptake. Since phosphate has a low diffusion coefficient, a rapid, concentrated event would be more effective than a long-term, lower level of exudation. The phenomenon could be limited by developmental constraints, related to meristem activity (Skene et al. 1998a,b). Johnson et al. (1996a,b) have examined the carbon economy of cluster roots in terms of exudation and enzyme activities.

The effects of reductants, chelating agents and protons on the rhizosphere of these plants are apparent. However, in nature plants live in communities, and the physiology of an individual plant has an impact on that community. Read (1996) has raised this issue in terms of coexistence of taxonomically distinct species in Australian heathlands. Cluster roots increase the availability of ions other than phosphate for neighbouring plants. There is evidence of manganese levels in plants without cluster roots becoming higher if grown alongside plants with cluster roots (Gardner & Boundy 1983). We can envisage a number of possible effects at a community level. For instance, is the species composition of a plant community affected by the presence of cluster roots, and in what ways? Could elevated levels of manganese and other metals enter the food chain or alter community composition in this way? Will changes in climate alter patterns of exudation, and if so, how will this impact on the plant communities involved?

Root patterns in a heterogeneous environment

Initiation of cluster roots occurs in response to internal rather than external levels of phosphate (Marschner et al. 1987). Further thought leads to the conclusion that their role in enhancing availability of nutrients through chelation and acidification in the rhizosphere means that previously unavailable phosphate becomes available, and so a cluster root is not dependent on the presence of nutrient-rich patches for its function. However, cluster roots do occur more in nutrient-rich patches, and Skene et al. (1996) have suggested that this is because lateral roots proliferate in such patches. Cluster roots occur at given intervals along lateral roots (Skene et al. 1996), and so in regions of lateral root proliferation more cluster roots will develop. Data from Lamont et al. (1974), for root and cluster root distribution of Leucadendron laureolum growing under natural conditions in South Africa, were used in a correlation analysis to examine the relationship between lateral root length and cluster root number in this heterogeneous environment (Skene 1997). The relationship, described by y = 0.118x + 0.691 (r2 = 99.9, P < 0.001), showed a strong correlation between lateral root occurrence and cluster root development. The predictive certainty of where the next cluster will occur also offers a unique insight into another great challenge to modern plant ecology, namely, why do roots occur where they do? Cluster roots are also well adapted for nutrient absorption, with large surface areas and soil volumes explored (Raven & Sprent 1993), and a higher rate of uptake than lateral roots. Dinkelaker et al. (1995) have suggested that there could be a period of intensive absorption following exudation.

Whether all cluster roots are identical functional units remains to be tested. Questions relating to classical competition theory and whether there is evidence of differential manipulation of the environment in terms of differences in exudation can be tested by examining niche separation (Cowling et al. 1994). Certainly, cluster roots undergo an exudative burst, but is this identical, in terms of quantity and quality, for each species and could any differences relate to niche separation? Any observed differences would be of great interest in terms of functional diversity of cluster roots. Finally, the dramatic changes that occur in the rhizosphere of cluster roots may offer the ideal setting in which to examine responses of microbes to changing events, giving an insight into community response in the rhizosphere. Since the exudative burst occurs in a fixed space, in an area where the rootlets have stopped growing, then the region of the rhizosphere affected will not itself change throughout these events.

While nitrogen-fixing nodules make use of bacterial nitrogenase and the high diffusion coefficient of N2, and mycorrhizas rely on the fine hyphae and efficient uptake systems of the fungus, cluster roots make use of basic root development and morphology to enhance uptake. The four key characteristics of the cluster root, the clustering of rootlets, the exudative burst, the determinate development of rootlets and the ellipsoid shape of the rootlet, are all seen to be traceable to alterations in basic root development (Skene 1997).

In conclusion, cluster roots are the third great adaptation for improved nutrient uptake in plants, the others being nitrogen fixation in root nodules and mycorrhizas. The lack of ecological research into cluster roots is surprising, and must be addressed, both in terms of their roles in nutrient cycling and the resulting impact on community structure. Our comprehension of how plant communities work will be sadly incomplete without a concerted effort to understand the roles played by the cluster root.


The comments of two anonymous referees, of Dr Lindsay Haddon and of Dr Ian Sanders are gratefully acknowledged.

Received 5 February 1998revision accepted 2 July 1998