Printer Friendly

Proteoid root mats bind surface materials in Hawkesbury Sandstone biomantles.


The sandy topsoils derived from Hawkesbury Sandstone are usually considered to be prone to severe sheet erosion following bushfire, with predicted yields of up to 48 t of sediment per hectare (Atkinson 1984). Although extensive export of charcoal and debris from slopes, and movement of mineral soil over the width of a few litter dams occur with each fire cycle, sediment yield of the magnitude predicted often does not occur as anticipated. It seems that there is an unaccounted mechanism preventing widespread stripping of apparently unprotected sandy topsoils.

Observations of a site near Sydney, burnt in January 1994, indicated that dense root mats were extensive at the soil surface. These were found to be the proteoid root mats of Banksia serrata L. f. Previous research on proteoid root systems has focussed on their role in enhancing nutrient uptake in plants on impoverished soils in Western Australia. It has been shown that the root hairs of proteoid roots adhere to sand grains, rocks, charcoal, humus particles, and raw litter (Lamont 1982). The extensive surface distribution of proteoid root mats observed following the January 1994 fires, and their known adherence to soil particles in low nutrient soils, suggested that proteoid root mats may provide a mechanism for limiting stripping of Hawkesbury Sandstone biomantles (sensu Johnson 1990). The hypothesis tested here was that proteoid root mats bind surface materials in Hawkesbury Sandstone biomantles.

Proteoid roots are dense proliferations of rootlets that occur on the root systems of nearly all genera in the Proteaceae (Lamont 1993). The specific morphology of proteoid roots varies with different species. In the genus Banksia they occur as compound proteoid roots which form a dense mat at the soil surface (Lamont 1984).

The areal extent of the root mat associated with any one Banksia serrata L. f. is up to 20 times the extent of the plant canopy (Jeffrey 1967). Where B. serrata L. f. trees are closely spaced, the proteoid root mat may be continuous through the soil and it is impossible to define the extent of the root mat associated with individual plants. Proteoid roots have been observed in the decomposing organic A0 horizon of soils and in the Al horizon to a depth of 15 cm in zones of nutrient concentration (Lamont and McComb 1974; Dell et al. 1980; Low and Lamont 1990; Lamont 1993).

Proteoid roots grow mainly from lateral roots. As a lateral root elongates, a sequence of proteoid roots may form along its length. Rootlets arise in longitudinal rows along the axis of the proteoid root (Purnell 1960) (Fig. 1). Mature rootlets are covered in root hairs that are thought to function like mycorrhizae, and to have a competitive advantage due to their greater tolerance to low soil moisture and their rapid development in response to moisture (Lamont 1982, 1993). Proteoid roots are known to excrete large amounts of sticky mucilage (Dell et al. 1980; Lamont 1993). Previous research has shown that proteoid roots are short-lived, seasonal structures. They shed their root hairs, the rootlet decomposes, and the root axis becomes woody within 12 months of initiation (Purnell 1960; Lamont 1972).


Materials and methods

Study sites

A site typical of Hawkesbury Sandstone hillslopes with respect to geology, topography, vegetation, and soils was selected at West Pymble in an area of Kuring-gai Council bushland reserve adjacent to Lane Cove National Park. The site is a gently sloping benched hillslope with alternating low cliffs formed by outcropping sandstone bedrock and regolith covered benches (Chapman and Murphy 1989). Soils are discontinuous due to extensive bedrock outcrops and rarely exceed 70 cm in depth. Soil types alternate between uniform coarse-textured soils and duplex soils (in the sense of Northcote 1979) formed by the downslope movement of coarse material over in situ clay material derived from shale or argillaceous sandstone. All soils have a well-developed biomantle. Vegetation is an open eucalypt woodland dominated by Eucalyptus piperita, E. gummifera (Sol. ex Gaertner) Hochr., and Banksia serrata L. f. with a shrub understorey. Distinctive understorey species are Xanthorrhoea media, Persoonia levis, Acacia terminalis, B. ericifolia L. f., Angophora hispida (Smith) Blaxell, and Leptospermum attenuatum. This site had been burnt i year previously by a wildfire (January 1994).

A control site with comparable topography, soils, and vegetation was selected at Westleigh. The Westleigh site was severely burnt in a wildfire in 1976 and then burnt in a fuel reduction burn in 1983. The Westleigh site was used to confirm the presence of proteoid roots and to assess the longer-term effect of fire on proteoid root mats.

Plotting the surface distribution of proteoid root mats

A 50 in by 60 in base map of the hillslope at the West Pymble site was prepared by surveying the site with a dumpy level. Rock outcrops, cliff-lines, and all boulders [is greater than] 50 cm across were plotted. The surface distribution of mineral soil, accumulated leaf litter (a year since fire had consumed standing leaf litter), and proteoid root mats was also plotted. In most cases, natural boundaries between soil materials were clearly distinguished and the surface clearly dominated by one material or another. The position of B. serrata L. f. trees was plotted and an average value for the spacing of B. serrata L. f. plants across the site calculated.

Collection of soil samples

Undisturbed soil samples were collected in Kubiena tins (90 by 90 by 70 mm) from both sites as part of a larger study that included analysis of proteoid root distribution in relation to geomorphic processes. For this reason the samples were collected from each of 6 identified micro-geomorphic elements. Three replicates were collected from each of the elements at both sites, making a total of 36 soil samples.

Treatment of soil samples

Soil samples were divided into 2 sets so that replicates from each type of micro-geomorphic element could be examined in 2 ways.

The first set of samples was air-dried and then horizontally sectioned within the Kubiena tin at 10-mm intervals. Each 10-mm layer of material was removed using a scalpel and brush. Following removal of each layer, a wooden block was placed beneath the tin and the tin pushed down to bring the sample surface flush with the top of the tin. Each new surface was vacuumed free of loose debris using a pipette attached to a vacuum line. This allowed examination of relatively undisturbed material under a binocular microscope with up to 80x magnification. Soil fabrics were described using the key developed by Humphreys (1985). Samples of soil fabrics containing proteoid roots were set aside for further observation and photography under a scanning electron microscope.

Preparation of carbowax samples

The second set of soil samples was oven-dried at 40 [degrees] C for 2 days, then impregnated with Carbowax 4000. Complete penetration of the sample required keeping the wax molten within the sample for 24 h and topping up the Carbowax as necessary. Once set, the samples were vertically sectioned into blocks 20 mm thick using a kerosene-lubricated diamond saw. Faces of the blocks were etched clean using gently running water. These vertical soil sections were compared with descriptions of the horizontally sectioned samples.

Preparation for scanning electron microscopy

Selected samples of the soil fabrics containing proteoid roots were attached to 10- and/or 28-mm stubs and painted around the base with colloidal silver to improve conductance. Using a sputter coater, each sample was then double-coated to achieve a gold coating thickness of approximately 20 nm. Each stub was examined and photographed under a scanning electron microscope at up to 700x magnification.


Surface distribution of proteoid root mats

On average, there was 1 mature Banksia serrata tree L. f. per 31.25 [m.sup.2]; 61% of B. serrata L. f. trees grew within 2 in of a low cliff-line or large boulder ([is greater than] 2 m long). Plotting the distribution of surface materials revealed that proteoid root mats occurred in patches across the entire site. Fig. 2 shows that surface proteoid root mats were concentrated in some areas and sparse in others. Of the total surface area of the West Pymble site, 39% was covered by mineral soil, 27% by rock outcrop, 24% by leaf litter, and 10% by proteoid root mat. Concentrations of surface proteoid root mats occurred: (i) adjacent to rock surfaces forming conspicuous sheaths around exposed rocks; (ii) in the apex of the depositional fans that formed at the base of joint crevices; and (iii) in the erosion zone of litter dam-microterrace features as described by Mitchell and Humphreys (1987). The orientation of proteoid root mat exposure closely followed the orientation of litter dam-microterrace features.


Subsurface distribution of proteoid root mats

Proteoid roots occurred in 29 of the 36 (81%) soil samples collected. Sectioning of the soil samples suggested that proteoid roots were much more widespread than indicated by their surface distribution. The proteoid root mats were concentrated in the uppermost 40 mm portion of the soil profile and were often overlain by a layer 1-5 mm thick of unconsolidated soil material, consisting of a combination of single-grain quartz sand, comminuted leaf litter, and charcoal fragments.

Soil fabrics associated with proteoid root systems

Four types of soil fabric containing proteoid roots of increasing density and fineness of material were identified.

Fabric Type 1: single-grain fabric

The roots in fabric Type I were sometimes charred and were characterised by an open lattice structure with no root hairs evident (Fig. 3). Rootlets remained attached to the proteoid root axis and were woody in appearance. Single grains of medium-to-coarse sand ([is greater than] 250 [micro]m) and charcoal lay within the lattice structure but were not bound to it in any way. Fig. 3 shows how individual particles were trapped by the rootlets. Fabric Type 1 occurred from the soil surface to a depth of 5 mm and was not overlain by any other soil material. It was, however, often associated with a coherent crust of fine particles [is less than] 2 [micro]m.


Fabric Type 2: organic fabric

Fabric Type 2 was characterised by densely hairy, cream-coloured roots (Fig. 4). Proteoid roots grew upward into overlying leaf litter and between flat-lying leaves. This bound leaves together imparting a laminated appearance to the root mat. Inwashed material lying between the leaves, such as comminuted leaf litter, charcoal fragments, single grain quartz sand, and fine mineral material was also trapped within the laminated structure. Fig. 4 shows the prolific root hairs adhering to the surface of the organic material. This material occurred from the soil surface to a depth of 30 mm at the unburnt site only and was overlain by loose leaf litter.


Fabric Type 3: porous-dense grain support fabric

Fabric Type 3 appeared curly and was characterised by the presence of discrete aggregates (Fig. 5). Rootlets were creamy-to-rusty coloured and retained root hairs. The root hairs adhered to comminuted leaf litter, charcoal fragments up to 5 mm long, and sand grains mostly in the size range 177-2000 [micro]m. Fig. 5a shows a quartz grain ~400 [micro]m across embraced by a rootlet and root hairs. Irregularly shaped aggregates ranging from 100 to 2000 [micro]m in diameter occurred. These were composed of fine quartz sand and finer materials [is less than] 2 [micro]. These aggregates were physically attached to the root hairs and appeared to have aggregated around the root hair (Fig. 5d). In spite of its porous appearance, this material was extremely coherent. Voids up to 12 mm across and maculae (sensu Humphreys 1994) were present. This fabric type occurred from a depth of 0 to 35 mm and was overlain by a combination of leaf litter, charcoal fragments, and single-grain quartz sands or fabric Types 1 or 2.


Fabric Type 4: dense grain support fabric-plasma support fabric

Fabric Type 4 had a concrete-like appearance and was strongly coherent. Root hairs were attached to the rootlets but were difficult to identify due to the contiguous coating of amorphous, fine material (Fig. 6). Root hairs appeared to be attached to individual grains ranging in size from 88 to 710 [micro]. The spaces around individual grains were almost completely infilled with amorphous, fine material [is less than] 2 [micro], probably a combination of charcoal, clays, and iron oxide. As with fabric Type 3, voids up to 12 mm across and patches of light yellow and light brown material occurred. Fabric Type 4 occurred from a depth of 5 to 40 mm and was overlain by a combination of leaf litter, charcoal fragments, and single-grain quartz sands or fabric Types 1 and 3.


All 4 fabric types were observed at the unburnt site, whereas only fabric Types 1, 3, and 4 were observed at the burnt site. Persistence of proteoid roots observed during this study was found to be highly variable. In some cases, the rootlets and root hairs were retained for at least 18 months. The presence of buried, charred root mats at the unburnt site suggests that some structural components of the root mat persist for up to 10 years (when, according to the recorded fire history, the site was last burnt).


The widespread occurrence of a persistent, coherent root mat which differentially binds soil particles in the uppermost portion of the mobile biomantle has 2 important implications. Firstly, the positioning, stability, and texture of the proteoid root mat mitigates the effectiveness of rainwash. Secondly, the root mat potentially contributes to the development of texture contrast soils.

Mitigation of rainwash

The position of the root mat high within the soil profile provides a physical barrier to the operation of erosive processes. At the unburnt site, proteoid roots were observed growing upward into the base of the leaf litter, forming an interlocking mat of leaves and binding it to the mineral soil surface. This has important implications for the downslope movement of both leaf litter and mineral soil. It is inferred from the evidence of stable binding mechanisms that the root mat protected the mineral soil from the action of rainsplash and rainwash, and trapped sediment washed into the leaf litter from upslope. This might have the effect over time of building up the soil profile by stabilising sediment.

At the burnt site the surface at the time of the fire was indicated by charring of the root mat. Fabric Type I remained coherent and was able to retain medium-to-coarse material even where the root mat had been damaged by fire and exposed to water erosion. The presence of unconsolidated material including charcoal fragments overlying fabric Types 3 and 4 indicated that some post-fire erosion had taken place on the slope. The stability of the root mat, however, provided a lower limit to which erosive processes could extend. This raises the question of the role of proteoid roots in re-distributing erosive processes across the slope, which is addressed elsewhere (Gould 1998).

The development of plasma support fabric associated with proteoid roots within the otherwise grain support fabric topsoil also suggests that proteoid roots reduce the effectiveness of rainwash. In their hypothesis of the pedological significance of combined bioturbation and rainwash, Humphreys and Mitchell (1983) envisaged a continuum of effects depending on the rate of rainwash, ground cover characteristics, and slope. Where rainfall is of sufficient intensity, it is expected to export fine material from the site, leading to the development of a coarser A horizon. However, where rainfall is of insufficient intensity, or where slope or groundcover characteristics mitigate the effect of rainwash, fine material would be re-incorporated into the soil profile. If the above hypothesis is correct, the presence of fabric Type 4 in association with maculae suggests that proteoid roots can reduce the effectiveness of rainwash and cause the re-incorporation of fine material deposited at the soil surface by bioturbation. Proteoid roots and bioturbation appear to operate in concert to produce plasma support fabric immediately below the soil surface.

Development of texture contrast

Proteoid roots do not bind all soil materials equally. Fabric analysis demonstrated binding mechanisms of differing persistence. This study confirmed that the first structural component of the root mat to break down is the root hairs. The most persistent binding mechanism was the rootlets. This structural component of the root mat can persist for years even following fire. The soil material most persistently bound was medium-to-coarse sand (177-2000 [micro]m). The differing persistence of binding mechanisms has substantial implications for the development of texture contrast soils. A model of the differential persistence of proteoid roots on soil particles of different size is proposed in Fig. 7.


Researchers at Macquarie University have demonstrated the role of bioturbation in the formation of texture contrast soils in the Sydney Basin (Humphreys and Mitchell 1983; Humphreys 1994; Paton et al. 1995). Within the model for the development of texture contrast soils, differential mobility of materials in the biomantle is due to preferential selection and deposition of fine material at the surface by soil fauna, and preferential removal of the finer components by rainwash. The results presented here demonstrate the possible operation of an additional process. In addition to the winnowing effect of bioturbation and rainwash proposed by Humphreys and Mitchell (1983), proteoid root mats actively retain medium-to-coarse material.

The arguments for proteoid roots mitigating the effectiveness of rainwash and contributing to the development of texture contrast initially appear contradictory. However, if the biomantle is envisaged as a series of circulating cells in which particles are removed at the surface more slowly with increasing size, then both processes can be considered to operate over different time scales. The actions of bioturbation, fire, and proteoid roots operate in concert over time to cause the coarsening of the biomantle.


It is concluded that proteoid root mats bind the surface materials in Hawkesbury Sandstone biomantles at the scale of individual soil particles and that they are sufficiently persistent and widespread that they may account for the stability of soil materials that appear to be unprotected from erosional forces. Proteoid root mats can be thought of as an organic form of geotextile. The laminated leaf layer acts to trap inwashed material which then becomes stabilised in the root mat. In this way, the root mat may be actively building the soil up. In addition, the action of proteoid roots in concert with bioturbation and fire probably contributes over time to the development of texture contrast soils in Hawkesbury Sandstone environments. The distribution of proteoid roots in relation to geomorphic processes and their role in stabilising Hawkesbury Sandstone biomantles at the scale of the slope are examined in more detail elsewhere (Gould 1998).


The author would like to acknowledge Peter Mitchell for supervision of the Honours project which forms the basis of this paper; Janet Eddy for her generous assistance in identification of proteiod roots, field work techniques, and preparation of Carbowax samples; Di Ward for assistance with soil fabric analysis; Coral Gillespie for assistance with preparation of SEM samples and use of equipment; and the School of Earth Sciences at Macquarie University for financial assistance.


Atkinson, G. (1984). Erosion damage following bushfires. Journal of Soil Conservation 40/41, 4-9.

Chapman, G. A., and Murphy, C. L. (1989). `Soil Landscapes of the Sydney 1:100000 Sheet.' (Soil Conservation Service of NSW: Sydney.)

Dell, B., Kuo, J., and Thomson, G. J. (1980). Development of proteoid roots in Hakea obliqua R. Br. (Proteaceae) grown in water culture. Australian Journal of Botany 28, 27-37.

Gould, S. F. (1998). Proteoid root mats stabilise Hawkesbury Sandstone biomantles following fire. Australian Journal of Soil Research 36, 1033-43.

Humphreys, G. S. (1985). Bioturbation, rainwash and texture contrast soils. PhD Thesis, School of Earth Sciences, Macquarie University.

Humphreys, G. S. (1994). Bioturbation, biofabrics and the biomantle: an example from the Sydney Basin. In `Soil Micromorphology: Studies in Management and Genesis'. (Eds A. J. Ringrose-Voase and G. S. Humphreys.) pp. 421-36. (Elsevier: Amsterdam.)

Humphreys, G. S., and Mitchell, P. B. (1983). A preliminary assessment of the role of bioturbation and rainwash on sandstone hillslopes in the Sydney Basin. In `Aspects of Australian Sandstone Landscapes'. (Eds R. W. Young and G. C. Nanson.) pp. 66-80. Australian and New Zealand Geomorphology Group Special Publication.

Jeffrey, D. W. (1967). Phosphate nutrition of Australian heath plants: I. The importance of proteoid roots in Banksia (Proteaceae). Australian Journal of Botany 15, 403-11.

Johnson, D. L. (1990). Biomantle evolution and the redistribution of earth materials and artifacts. Soil Science 149 (2), 84-102.

Lamont, B. B. (1972). The morphology and anatomy of proteoid roots in the genus Hakea. Australian Journal of Botany 20, 155-74.

Lamont, B. B. (1982). Mechanisms for enhancing nutrient uptake in plants with particular reference to mediterranean South Africa and Western Australia. The Botanical Review 48, 597-678.

Lamont, B. B. (1984). Specialised modes of nutrition. In `Kwongan: Plant Life of the Sandplain'. (Eds J. S. Pate and J. S. Beard.) pp. 126-145. (University of Western Australia Press: Perth.)

Lamont, B. B. (1993). Why are hairy root clusters so abundant in the most nutrient-impoverished soils of Australia? Plant and Soil 155/156, 269-72.

Lamont, B. B., and McComb, A. J. (1974). Soil microorganisms and the formation of proteoid roots. Australian Journal of Botany 22, 681-8.

Low, A. B., and Lamont, B. B. (1990). Aerial and below-ground phytomass of Banksia scrub-heath at Eneabba, south-western Australia. Australian Journal of Botany 38, 351-9.

Mitchell, P. B., and Humphreys, G. S. (1987). Litter dams and microterraces formed on hillslopes subject to rainwash in the Sydney Basin, Australia. Geoderma 39, 331-57.

Northcote, K. H. (1979) `A Factual Key for the Recognition of Australian Soils.' 4th Edn. (Rellim Technical Publications: Glenside, S. Aust.)

Paton, T. R., Humphreys, G. S., and Mitchell, P. B. (1995). `Soils: A New Global View.' (UCL Press: London.)

Purnell, H. M. (1960). Studies of the Family Proteaceae: I. Anatomy and morphology of the roots of some Victorian species. Australian Journal of Botany 8, 38-50.

Manuscript received 12 January 1998, accepted 23 June 1998

Susan F. Gould

PO Box 478, Weipa, Qld 4874, Australia; email:
COPYRIGHT 1998 CSIRO Publishing
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1998 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Gould, Susan F.
Publication:Australian Journal of Soil Research
Date:Nov 1, 1998
Previous Article:Procedures for characterising denitrification rates in a wastewater-irrigated forest soil.
Next Article:Effects of time and temperature on the sorption of cadmium, zinc, cobalt, and nickel by a soil.

Terms of use | Copyright © 2017 Farlex, Inc. | Feedback | For webmasters