Agroforestry Systems 45: 303–341, 1999.

 1999 Kluwer Academic Publishers. Printed in the Netherlands.

Application of the ecosystem mimic concept to the

species-rich Banksia woodlands of Western Australia

J. S. PATE* and T. L. BELL

Department of Botany, The University of Western Australia, Nedlands 6907, Australia (*Author
for correspondence: [email protected])

Key words: life form, net primary productivity, phenology, rooting patterns, trophic speciali-
sations, water relations

Abstract. This article describes the structure and functioning of a natural Banksia woodland at
Moora, Western Australia. Species are first grouped in terms of growth form, root morphology,
phenology and nutrient acquisition strategy. Above- and belowground standing biomass of a
woodland is measured and its net annual primary production per unit rainfall compared with that
of adjacent crops and plantings of the tree tagasaste. Information on seasonal water use and
nutrient cycling in the dominant tree species Banksia prionotes is used to highlight the pivotal
role of deep rooted summer growing trees in the maintenance of sustainability of the system.
The article then addresses how one might select species mixtures as functionally effective ana-
logues of the woodland. Assuming the mimic system replaces cleared virgin woodland not pre-
viously subject to runoff of water and nutrients from agriculture, a selection procedure would
incorporate native flora representing (a) summer-growing deep-rooted and winter-growing
shallow-rooted trees and shrubs, (b) herbaceous ground cover species, (c) fire resistant and fire
sensitive species, and (d) a range of complementary nutrient uptake strategies. Assuming the
mimic is designed to rehabilitate agricultural land experiencing rising water tables and nitrate
pollution of ground water, incorporation of fast growing deep rooted exotic trees or herbaceous
perennials is recommended alongside conventional annual crops or pastures, with appropriate
nutrient stripping through removal of biomass. Difficulties in this context are scale of planting
required and current lack of profitable incentives for planting and maintenance of perennials.

1. Introduction

This article sets out to describe, and then suggest how one might proceed to
mimic, a complex functionally sustainable natural ecosystem. The system
chosen is south western Australian Banksia woodland, embodying some of
the most oligotrophic soils and exacting climatic conditions experienced
anywhere on the continent. Banksia woodlands, defined as containing at least
one dominant species of Banksia, other trees at lesser density and an under-
storey of predominantly woody shrubs, comprise one of the major vegeta-
tional types of the deep sands of the Swan Coastal Plain and Dandaragan
Plateau. They are more or less confined to a narrow strip of land 15 to 90 km
wide extending for 400 km between Jurien Bay in the north to Busselton in
the south (Figure 1a). Much of the biome (approximately 61%) has been
cleared for agriculture and urban and semi-rural development or exploited
for natural resources such as sand and limestone (Gozzard and Mouritz, 1989)
and timber (Hopkins, 1989). Nevertheless, large tracts of Banksia woodlands
304

Figure 1. (a) Distribution of Banksia woodlands in southwest Australia with details of mean
annual rainfall, maximum and minimum temperatures for four sites; (b) Moora, (c) Gingin,
(d) Kings Park, Perth and (e) Bunbury. Source: Map (a) redrawn from Beard (1989), weather
information in (b) to (e) from the Bureau of Meteorology, Perth.
 305

and associated open heaths remain in relatively pristine state on Crown Lands
or conservation reserves (Beard, 1989).
Several species of Banksia are characteristic dominants of the woodlands,
the commonest of these being B. attenuata, B. menziesii and B. ilicifolia and,
in low-lying wetter areas, B. littoralis. Accompanying tree species include
Actinostrobus arenarius, Allocasuarina fraseriana, Banksia burdettii, B.
prionotes, Corymbia calophylla, Eucalyptus marginata, E. todtiana, Nuytsia
floribunda and Xylomelum angustifolium (Beard, 1989; Dodd and Griffin,
1989). The species-rich understorey is dominated by sclerophyllous woody
species from the families Proteaceae, Myrtaceae, Fabaceae and Epacridaceae.
The equally rich herbaceous ground flora includes members of the
Restionaceae, Cyperaceae, Orchidaceae, Haemodoraceae and Anthericaceae
(Dodd and Griffin, 1989). Life forms typically encompass relatively few
species of tall trees (4–8 m in height), a taxonomically diverse mid-storey of
woody shrubs (0.2–2.5 m) and a herbaceous ground cover of hemicrypto-
phytes, geophytes, helophytes, and epiphytes (Pate et al., 1984). For example,
in a broad survey conducted by Speck (1952) in Banksia woodlands of the
Swan Coastal Plain near Perth, 28 species of trees and tall shrubs, 130 species
of low shrubs and 182 species of herbaceous ground layer plants were listed
as common components.
Soils of Banksia woodlands variously comprise well-drained and highly
leached, yellow or white quartz sands of the inland Bassendean Dune Systems,
the shallow to deep yellow, iron-coated quartz sands of the Spearwood Dune
Systems or, in lower lying areas, poorly-drained organic matter-rich soils
bordering interdunal swamps and lakes (Bettenay, 1984; Semeniuk and
Glassford, 1989). All of these sands are exceptionally poor in nutrients and
base-exchange capacity, typically in the pH range of 4.5–6.0, and highly
deficient in clay components. For example, in Bassendean sands of Banksia
woodlands near Gingin, 75 km north of Perth, Bell et al. (1994) recorded
greatest concentrations of available nutrients within the top 20 cm of soil.
Below this, concentrations quickly declined until, at about 50 cm depth,
mineral contents were only marginally greater than those of the parent sands
of the ecosystem. Soil organic matter content in the top 20 cm was only 0.7%
and maximum levels of critically important macronutrients in this zone
were 1 µg NO3––N g dry weight soil–1, 12 µg NH4+–N, 2–5 µg total P and
40 µg K.
It should be noted that similar soils, loosely termed ‘Banksia sands’, occur
in discrete pockets from the southwest corner of the State, eastwards along
the coast to Albany and thereafter surrounding Esperance. Vegetation in these
localities is also dominated by Proteaceae, but mostly in shrubland rather than
woodland format. We will restrict our attention here to Banksia woodlands
as detailed in Figure 1a.
The climate experienced by most of the Botanical Province of the south-
west corner of Australia is classic ‘mediterranean-type’, with hot dry summers
alternating with cool wet winters. The area occupied by Banksia woodlands
306

is classified as moderate to severe ‘dry mediterranean’ (Beard, 1984, 1989),

with a dry summer season spanning five to six months of the year (November
to April) when no useful rain is likely to fall. Winter rainfall commences in
May and peaks in June, July and early August, with rain-bearing cyclonic
systems penetrating first from the northwest and then more widely from a
series of wet depressions from the south west (Gardner, 1944). Effective
rainfall decreases sharply during late August and September but attenuates
later in the south than north of the biome. Figure 1(b) to (d) illustrates rainfall
data and maximum and minimum temperatures for the last 30–100 years at
four sites (i.e. Moora, Gingin, Kings Park and Bunbury) spanning the latitu-
dinal extent of the biome.
Based on the hypothesis that soil texture and moisture relations play a more
important role than nutrient availability in shaping the physiognomy, life form,
and taxonomic spread within SW Australian ecosystems (Beard, 1984),
Banksia woodlands on relatively shallow sands overlying ironstone gravel are
almost invariably restricted to areas of annual rainfall of 600 to 750 mm,
whereas those on deeper leached sands extend more broadly across areas of
annual rainfall ranging from 500 to 900 mm (Beard, 1984). In deep sand
ecosystems further south in regions of even greater rainfall (900 to 1400 mm),
eucalypt-dominated forests typically replace Banksia woodlands, while at the
other extreme, in areas of ‘Banksia sands’ north and south of the State
receiving less than 400 mm rainfall, low density Banksia woodlands inter-
grade with, and are finally replaced by scrub heath, albeit still dominated by
proteaceous species (see Hobbs and O’Connor, this volume).
Seasonal changes in soil moisture under Banksia woodlands (e.g. see Dodd
et al., 1984; Bell et al., 1994) show protracted minima during summer months,
January to April, increasing gradually with the onset of winter rains in May
and June, and remaining transiently close to or at field capacity (5% by weight)
during winter months of July and August. After this, the soil dries out
progressively from September to November and remains very dry until the
following autumn. Large variations can occur between seasons depending on
when and in what quantity rainfall events occur. For example, at our prin-
cipal study site in northern Banksia woodlands at Moora, a whole year passed
in the exceptionally dry winter of 1997 without the soil profile becoming fully
wetted from the surface down to the groundwater at 5 m depth. Not surpris-
ingly, seasonal shoot extension of trees during the following summer proved
to be minimal (A. M. Grigg, M. J. Unkovich and J. S. Pate, unpubl. data).

2. Morphological and functional characteristics of Banksia woodland

In the following sections we assemble information on the phenologies and

morphological and physiological characteristics exhibited by floristic elements
of typical Banksia woodlands, using as examples data from Melaleuca Park
and Gnangara, 40–50 km north of Perth (Dodd et al., 1984; Farrington et al.,
 307

1989; Farrington and Bartle, 1991), Kings Park, Perth (Stewart et al., 1993;
Pate et al., 1993) and Moora, 250 km north of Perth (T. E. Dawson, T. J.
Hatton, D. J. Arthur, M. J. Unkovich, A. M. Grigg, T. L. Bell and J. S. Pate,
unpubl. data). Banksia spp. and other proteaceous species collectively make
up to 70% or more of the ground cover at these sites, so special attention
will be directed at water and nutrient relationships of members of this family.
A later section of the article presents information relating to one particularly
well researched species, Banksia prionotes (see review of Pate et al., 1998).

2.1. Growth and life forms, phenologies of growth and flowering

Table 1 lists the common species encountered at the Moora and Kings Park
sites, according to growth and life form and the times of the year when shoot
extension and flowering would normally occur. The majority of understorey
species at Moora are woody shrub species (57% of species) alongside fewer
species of herbaceous perennials (24% of species). At Kings Park, the pro-
portion of herbaceous perennial species (43%) is slightly greater than woody
shrubs (37%). Both sites contain relatively small percentages of their floras
as ephemerals (8% of species at Moora and 9% of species at Kings Park) and
trees (6% of species at Moora and 7% of species at Kings Park). Dominance
of the families Proteaceae (24% of species at Moora and 13% of species at
Kings Park), Myrtaceae (19% of species at Moora and 4% of species at Kings
Park) and Fabaceae (5% of species at Moora and 9% of species at Kings Park)
is clearly evident amongst shrub and tree components. A phenology involving
shoot extension in winter accompanied by flowering in winter or early spring
(hereafter referred to as winter-active species) is exhibited by a preponder-
ance of species (71% at Moora and 81% at Kings Park) compared with those
showing summer shoot extension and summer or early autumn flowering
(hereafter termed summer-active species), for which the relevant proportions
are 29% at Moora and 19% at Kings Park. It must be remembered that the
term ‘summer-active’ refers solely to timing of shoot extension and flowering
and that species of this ilk are likely to be ‘winter-active’ in terms of nutrient
uptake in having their feeding roots located mostly in top layers of soil.

2.2. Responses to fire

Banksia woodlands are frequently subjected to fire and, in common with other
south western Australian ecosystems, contain widely ranging mixtures of
‘resprouter’ species surviving fire by sprouting from root stocks or trunks,
and ‘obligate seeders’ or ‘non-sprouter’ species succumbing to fire and
recruiting thereafter exclusively from seed (Bell et al., 1984). As shown by
Pate et al. (1990) and Bell and Pate (1996), resprouters are typically long-
lived, poorly active in reproduction, of slow growth during seedling stages,
late in achievement of adulthood and exhibit shoot:root dry weight ratios in
the range of 0.6 to 2.5. Seeders by contrast are short-lived, fast growing, highly
 308
Table 1. Life and growth forms, shoot and flowering phenology, fire response and rooting type of common species encountered in two Banksia woodland
sites at (a) Moora and (b) Kings Park.

Life and Family *Species Shoot Flowering Fire Rooting

growth form extension phenology response type

(a) MOORA
Ephemerals Asteraceae *Hypochaeris glabra L. W W/Sp S 1
*Podotheca gnaphalioides Grah. W Sp S 1
*Ursinia anthemoides (L.) Gaertn. W Sp S 1
*Waitzia acuminata Steetz in Lehm. W Sp S 1
Geraniaceae *Erodium cygnorum Nees W W S 1
Gyrostemonaceae *Gyrostemon ramulosus Desf. W W/Sp S 2
Herbaceous Chloanthaceae *Pityrodia bartlingii (Lehm.) Benth. W Sp/S R 3
perennials *Pityrodia verbascina (F. Muell.) Benth. W W R 3
Cyperaceae *Caustis dioica R. Br. W Sp R 1
*Mesomelaena stygia (R. Br.) Nees W W R 1
Dasypogonaceae *Lomandra sp. W Sp R 1
Droseraceae *Drosera erythrorhiza Lindl. W W/Sp R 1
*Drosera sp. (pygmy) W W S 1
Goodeniaceae *Verrauxia reinwardtii (Vriese) Benth. Sp Sp/S R 3
Haemodoraceae *Anigozanthos humilis Lindl. W W/Sp R 1
*Conostylis setigera R. Br. W Sp R 1
Orchidaceae *Caladenia flava R. Br. W Sp R 1
*Lyperanthus nigricans R. Br. W Sp R 1
Phormiaceae *Dianella revoluta R. Br. S Sp R 1
Poaceae *Neurachne alopecuroidea R. Br. W W/Sp R 1
Restionaceae *Lepidobolus preissianus Nees W W/Sp R 1
Rubiaceae *Opercularia vaginata Labill. S Sp S 2
Rutaceae *Boronia sp. W W S 2
Stylidiaceae *Stylidium aff. repens R. Br. W W/Sp S 1
*Stylidium brunonianum ssp. brunonianum Benth. W/Sp Sp S 1
Parasites Lauraceae *Cassytha sp. –
Loranthaceae *Nuytsia floribunda (Labill.) R. Br. Ex Fenzl S S R 5
Orobanchaceae *Orobanche minor Sm. W Sp S –
Woody shrubs Casuarinaceae *Allocasuarina humilis (Otto & Dietr.) L. Johnson S Sp/S R 5
Dilleniaceae *Hibbertia acerosa (R. Br. Ex DC.) Benth. W Sp R 2
*Hibbertia pachyrrhiza Steudel Sp Sp/S S 2
*Hibbertia sp. W S 2
Epacridaceae *Astroloma xerophyllum (DC.) Sonder W W S 2
*Leucopogon polymorphus Sonder W W S 2
*Lysinema ciliatum R. Br. W W S 2
Mimosaceae *Acacia pulchella ssp. glaberrima Meissner S W S 4
Myrtaceae *Baeckea camphorosmae Endl. S Sp/S S 4
*Beaufortia elegans Schauer S S R 4
*Calothamnus sanguineus Labill. W Sp R 4
*Calytrix fraseri Cunn. Sp W/Sp R 2
*Calytrix sp. Sp Sp/S S 2
*Eremaea beaufortioides Benth. S Sp R 5
*Leptospermum erubescens Schauer S W/Sp R 5
*Leptospermum spinescens Endl. W Sp R 4
*Melaleuca scabra R. Br. W Sp/S R 5
*Melaleuca sp. (round fruit) W Sp/S R 4
*Pileanthus filifolius Meisn. W Sp/S S 1
*Verticordia acerosa Lindl. W Sp S 2
*Verticordia densiflora Lindl. W Sp S 2
*Verticordia sp. (large pink) W Sp S 2
Fabaceae *Gompholobium sp. (yellow flower) S Sp S 2
*Jacksonia aff. stricta S Sp R 5
*Jacksonia ulicifolia S Sp R 5
*Mirbelia spinosa Benth. W/Sp W S 2
Proteaceae *Adenanthos cygnorum Diels W Sp/S S 5
*Banksia leptophylla A. S. George S S S 5
*Conospermum incurvum Lindl. W W/Sp S 3

309
 310
Table 1. (Continued)

Life and Family *Species Shoot Flowering Fire Rooting

growth form extension phenology response type

*Conospermum stoechadis Endl. W W/Sp R 5

*Dryandra nivea (Labill.) R. Br. W W/Sp R 5
*Dryandra sessilis (Knight) Domin W W/Sp S 5
*Grevillea eriostachya ssp. eriostachya McGillivray W W/Sp S 5
*Hakea obliqua R. Br. S W/Sp S 5
*Hakea platysperma Hook. S W/Sp S 5
*Hakea trifurcata (Smith) R. Br. W/Sp Sp R 5
*Isopogon dubius (R. Br.) Druce S Sp R 5
*Isopogon scabriusculus Meissn. W/Sp Sp R 5
*Petrophile brevifolia Lindl. S Sp R 5
*Stirlingia latifolia (R. Br.) Steudel Sp/S Fire R 5
*Synaphea spinulosa ssp. spinulosa (Burm. F.) Merr. W W/Sp R 5
Rhamnaceae *Cryptandra pungens Steudel W W S 2
Thymelaeaceae *Pimelea leucantha Diels W/Sp W/Sp S 2
*Pimelea spectabilis Lindl. W Sp S 2
Trees Cupressaceae *Actinostrobus arenarius W S 5
Myrtaceae *Eucalyptus todtiana F. Muell. S S R 5
Proteaceae *Banksia attenuata R. Br. S Sp/S R 5
*Banksia menziesii R. Br. S S R 5
*Banksia prionotes Lindl. S S S 5
*Xylomelum angustifolium Kipp. et Meissn. S S R 5

(b) KINGS PARK

Ephemerals Apiaceae *Eryngium rostratum Cav. W/Sp W/Sp S 2
Asteraceae *Podolepis gracilis (Lehm.) Grah. W W/Sp S 1
*Podotheca angustifolia (Labill.) Less. W Sp S 1
*Podotheca chrysantha (Steetz) Benth. W Sp/S S 1
 *Ursinia anthemoides (L.) Gaertn. W Sp S 1
*Waitzia suaveolens (Benth.) Druce W Sp/S S 1
Cyperaceae *Isolepis marginata (Thunb.) A. Dietr. W/Sp Sp S 1
Euphorbiaceae *Poranthera microphylla Brongn. W Sp S 1
Lobeliceae *Lobelia gibbosa Labill. W Sp/S S 1
*Lobelia tenuior R. Br. W Sp/S S 1
Herbaceous Amaranthaceae *Ptilotus polystachyus (Gaudich.) F. Muell. Sp W/Sp R 3
perennials Anthericaceae *Arthropodium preissii Endl. W/Sp Sp/S R 2
*Corynotheca micrantha (Lindl.) J. F. Mcbride W/Sp S R
*Sowerbaea laxiflora Lindl. W/Sp W/Sp R 1
*Thysanotus manglesianus Kunth W/Sp Sp R 1
Asteraceae *Helichrysum cordatum DC. Sp/S Sp/S R 1
Colchicaceae *Burchardia umbellata R. Br. W Sp R 1
Cyperaceae *Lepidosperma angustatum R. Br. W/Sp W/Sp R 1
*Lepidosperma leptostachyum Benth. W/Sp W/Sp R 1
*Lepidosperma scabrum Nees W/Sp W R 1
*Mesomelaena stygia (R. Br.) Nees W/Sp S R 1
*Schoenus curvifolius (R. Br.) Benth. W/Sp W/Sp S 1
*Schoenus grandiflorus (Nees) F. Muell. W/Sp W R 1
*Tetraria octandra (Nees) Kuek. W/Sp W/Sp 1
Dasypogonaceae *Calectasia cyanea R. Br. W/Sp W/Sp R 2
*Dasypogon bromeliifolius R. B. Sp/S Sp/S R 1
*Lomandra caespitosa (Benth.) Ewart W/Sp W/Sp 1
*Lomandra nigricans T. D. Macfarlane W/Sp W 1
*Lomandra preissii (Endl.) Ewart W/Sp W 1
Droseraceae *Drosera pallida Lindl. W W/Sp 1
Fabaceae *Hardenbergia comptoniana (Andrews) Benth. Sp W/Sp R 5
*Kennedia prostrata R. Br. Sp Sp S 5
Haemodoraceae *Anigozanthos humilis Lindl. W W/Sp S 1
*Anigozanthos manglesii D. Don W/Sp Sp S 1
*Conostylis aculeata R. Br. W/Sp Sp 1
*Haemodorum laxum R. Br. W/Sp Sp R 1

311
 312
Table 1. (Continued)

Life and Family *Species Shoot Flowering Fire Rooting

growth form extension phenology response type

*Haemodorum paniculatum Lindl. W/Sp Sp/S R 1

*Haemodorum spicatum R. Br. W/Sp Sp/S R 1
Iridaceae *Patersonia occidentalis R. Br. Sp Sp R 1
Orchidaceae *Caladenia deformis R. Br. W W R 1
*Caladenia flava R. Br. W Sp R 1
*Caladenia latifolia R. Br. W W/Sp R 1
*Caladenia longicauda Lindl. W W/Sp R 1
*Diuris longifolia R. Br. W Sp R 1
*Elythranthera brunonis (Endl.) A. S. George W Sp R 1
*Lyperanthus nigricans R. Br. W Sp R 1
*Pterostylis scabra var. robusta (Rogers) A. S. George W W R 1
Phormiaceae *Dianella divaricata R. Br. W/Sp Sp R 1
Restionaceae *Alexgeorgea nitens Sp W R 1
*Desmocladus fasciculatus Sp W/Sp R 1
*Desmocladus flexuosus Sp Sp R 1
*Hypolaena exsulca R. Br. Sp Sp/S R 1
*Lyginia barbata R. Br. Sp W/Sp R 1
Rubiaceae *Opercularia vaginata Labill. S Sp S 2
Stackhousiaceae *Tripterococcus brunonis Endl. W/Sp W/Sp S 1
Stylidiaceae *Stylidium brunonianum Benth. W/Sp Sp S 1
*Stylidium calcaratum R. Br. W Sp S 1
Violaceae *Hybanthus calycinus (DC. ex Ging.) F. Muell. W/Sp Sp S 2
Parasites Lauraceae *Cassytha glabella R. Br. S –
Loranthaceae *Amyema miquelii (Lehm. ex Miq.) Tieghem W/Sp S –
*Nuytsia floribunda (Labill.) R. Br. Ex Fenzl S S R 5
Orobanchaceae *Orobanche minor Sm. W Sp S 1
Woody shrubs Casuarinaceae *Allocasuarina humilis (Otto & Dietr.) L. Johnson S Sp/S R 5
Dilleniaceae *Hibbertia huegelii (Endl.) F. Muell. Sp S 2
*Hibbertia hypericoides (DC.) Benth Sp R 2
Epacridaceae *Astroloma macrocalyx Sond. W W R 2
*Astroloma pallidum R. Br. W W R 2
*Conostephium pendulum Benth. W W R 2
*Conostephuim preissii Sond. W W R 2
*Leucopogon propinquus R. Br. W W R 2
Euphorbiaceae *Phyllanthus calycinus Labill. W/Sp Sp R 2
Fabaceae *Bossiaea eriocarpa Benth. W/Sp R 3
*Daviesia divaricata Benth. W/Sp R 3
*Daviesia nudiflora Meiss. W R 3
*Daviesia triflora M. D. Crisp W/Sp R 3
*Gompholobium tomentosum Labill. Sp/S Sp/S S 2
*Hovea trisperma Benth. W/Sp Sp S 2
*Isotropis cuneifolia (Smith) Benth. ex B. D. Jackson W/Sp Sp R 2
Fabaceae *Jacksonia furcellata (Bonpl.) DC. S Sp/S R 5
*Jacksonia sericea Benth. S S R 5
*Oxylobium capitatum Benth. W/Sp W/Sp R 2
Goodeniaceae *Dampiera linearis R. Br. W/Sp R 2
*Scaevola canescens Benth. W/Sp R 2
*Scaevola holosericea De Vriese Sp R 2
*Scaevola paludosa R. Br. Sp R 2
Mimosaceae *Acacia cochlearis (Labill.) H. L. Wendl. Sp 5
*Acacia pulchella R. Br. S W S 4
*Acacia stenoptera Benth. W/Sp R
Myrtaceae *Calytrix fraseri A. Cunn. W S R 2
*Hypocalymma robustum Endl. W Sp R 2
*Melaleuca acerosa Schau. W Sp R 2
Proteaceae *Adenanthos cygnorum Diels W Sp/S S 4
*Dryandra nivea (Labill.) R. Br. W/Sp R 5
*Dryandra sessilis (Knight) Domin W/Sp S 5

313
 314
Table 1. (Continued)

Life and Family *Species Shoot Flowering Fire Rooting

growth form extension phenology response type

*Grevillea vestita Endl. Sp/S W 5

*Hakea prostrata R. Br. S Sp R 5
*Persoonia saccata R. Br. W/Sp S
*Petrophile linearis R. Br. Sp/S Sp R 4
*Petrophile macrostachya R. Br. Sp/S Sp 4
*Stirlingia latifolia Endl. Sp/S Fire R 4
Rhamnaceae *Spyridium tridentatum (Steudel) Benth. S S 2
Rutaceae *Eriostemon spicatus A. Rich. W/Sp W/Sp 2
Xanthorrhoeaceae *Xanthorrhoea preissii Endl. Sp/S Sp R 1
Zamiaceae *Macrozamia riedlei (Fischer ex Gaudich.) C. Gardner Sp/S W/Sp R 5
Trees Casuarinaceae *Allocasuarina fraseriana (Miq.) L. Johnson S W/Sp R 5
Myrtaceae *Eucalyptus calophylla Lindl. S S R 5
*Eucalyptus marginata Donn ex Sm. S S R 5
Proteaceae *Banksia attenuata R. Br. S S R 5
*Banksia grandis Willd. S S R 5
*Banksia ilicifolia R. Br. S S R 5
*Banksia menziesii R. Br. S S R 5
*Banksia prionotes Lindl. S S S 5
Key and summary Number of species

Moora Kings Park

* – Introduced species 03 02
Shoot extension W– winter (June to August) 45 30
W/Sp – winter/spring (July to October) 05 31
Sp – spring (September to November) 04 09
Sp/S – spring/summer (October to December) 01 09
S – summer (December to February) 22 15
Flowering phenology W – winter (June to August) 11 15
W/Sp – winter/spring (July to October) 18 32
Sp – spring (September to November) 27 36
Sp/S – spring/summer (October to December) 11 14
S – summer (December to February) 07 13
Fire – fire stimulated flowering 01 01
Fire response R – resprouter 39 31
S – obligate seeder 38 71
Rooting types (see Figure 2) 1 – no obvious main root, shallow laterals 20 51
2 – shallow single main root, shallow laterals 19 26
3 – deep tap root, few laterals 04 05
4 – deep tap root, laterals throughout profile 06 05
5 – one or more deep tap roots, shallow laterals 27 20

315
316

reproductive and exhibit high shoot:root ratios in the range of four to 10.
Table 1 provides information on the respective fire response status of the
common species listed for the two Banksia woodland sites. Note the pre-
dominance of resprouters (70% of species) at the Kings Park site and better
representation of seeders (50% of species) at the Moora site.
Hot fires, which periodically consume all flammable biomass of Banksia
woodlands, have very dramatic effects on vegetation. Seeds of seeders and
some resprouters, and particularly of short-lived fire ephemerals, germinate
prolifically and exhibit rapid initial growth rates in post-fire environments
(see Pate et al., 1985). Fast growth is possible since water, light and limiting
nutrients such as phosphorus are at much greater availability after fire than
in fully vegetated long unburnt communities. Aboveground living biomass is
restored to pre-fire levels in the order of five to ten years, being at first
composed principally of replacement shoots of resprouters and then increas-
ingly of new shoots of recruiting seeders (Bell et al., 1984; Hobbs and Atkins,
1990). Short-lived understorey legumes provide substantial amounts of fixed
N at this stage (Pate and Unkovich, in press) and these inputs are highly
significant in replacing N losses to the atmosphere during fire. This fixed N
is first released from decomposing legume biomass and then cycles progres-
sively through that of other community members. However, with increasing
time after fire, the woodland returns to its supposedly basic P-limited condi-
tion (see Pate and Dell, 1984), whereupon success of mycorrhizal and other
agencies concerned principally with P acquisition becomes paramount to
growth and survival of species.

2.3. Rooting morphologies

Using the simplified scheme for categorising rooting morphologies shown in

Figure 2, the Moora and Kings Park sites provide examples of all categories
depicted (see entries of Table 1), but with all of their large tree and shrub
species possessing deeply penetrating tap roots (Root types 3–5). Such
phreatophytic species access deeply located soil water resources and, as seen
in Table 1, tend to be mostly summer-active in their shoot growth and
flowering. Conversely, shallow-rooted shrubs, herbaceous perennials and
ephemerals, geophytes and hemicryptophytes conform mostly to rooting
patterns 1 and 2, thus constraining their growth and reproduction to winter
and early spring. They survive the summer in a dormant drought-tolerant state,
induced rapidly once the upper soil profile in which their roots are located
has dried out to critical levels (see also Dodd et al., 1984; Dodd and Heddle,
1989; Specht et al., 1991).
Woody species of all categories develop crops of fine multi-branched roots
on the perennial parts of their root systems each autumn and these engage in
intense uptake of nutrients through winter and spring until the upper soil dries
in early summer (Pate and Jeschke, 1993). Feeding roots of this nature are
confined mostly to the upper 50 cm or so of the soil profile, where the bulk
 317

Figure 2. Schematic representation of five basic rooting patterns encountered in deep sands of
Banksia woodlands in south western Australia. Types 1 and 2 are typically shallow, reaching
depths of only 50 to 70 cm, whereas Types 3 to 5 invariably attain greater depths and reach the
water table if at accessible depth (dashed line).

of the predominantly organically-bound reserves of nutrients are located.

Interspecific competition for nutrients through various mycorrhizal-mediated
and non-mycorrhizal strategies is particularly intense at times when these
upper layers of the soils are suitably moist (Bell, 1995). According to studies
on seedling Proteaceae by Bowen and Pate (1991), seeder species typically
make unusually greater proportional investments into biomass of lateral
feeding roots, presumably to accommodate the large nutrient demands asso-
ciated with the fast growth rates and high reproductive outputs typical of
seeders. The corresponding situation for resprouters shows accentuated
biomass allocation into enlarged tap roots and lignotubers, and deposition
within these organs of substantial carbohydrate reserves. The latter are utilised
during resprouting after fire (Bowen and Pate, 1993; Bell et al., 1994).

2.4. Water relations of ecosystem and species

Studies at Gnangara by Dodd et al. (1984), Farrington et al. (1989) and

Farrington and Bartle (1991) provide evidence of annual fluctuations of the
order of 500–800 mm in water table depths of unconfined groundwater
systems typical of much of the higher rainfall regions of the Banksia woodland
biome. Minimum depth of water table is normally reached in September or
October, following percolation of spring rain through a soil profile already
318

fully recharged by the peak winter rains of June, July and August (see
Figure 3). The corresponding maximum then occurs in late autumn, when sand
several metres down the profile is still fully dried out and deep-rooted summer-
active species are drawing extensively from the capillary fringe of a retreating

Figure 3. (a) Seasonal variation in soil moisture content at Melaleuca Park, Western Australia
during drying (August to April) and recharge by winter rainfall (May to July), and (b) depth to
water table during the same period (August 1981 to September 1982). Source: Redrawn from
Dodd et al. (1984).
 319

water table. Judging from seasonal patterns of evapotranspiration of shrubs

such as Adenanthos cygnorum, Eremaea pauciflora and Stirlingia latifolia
studied by Dodd et al. (1984), and similar information for tree banksias (B.
attenuata and B. menziesii), A. cygnorum and groundflora components of
Banksia woodlands provided by Farrington et al. (1989), peak water loss to
the atmosphere typically occurs in late spring and early summer and corre-
sponding minima in the cooler and wetter months of June and July (Figure
4a). The study of Farrington et al. (1989), comparing yearly losses as total
evapotranspiration (transpiration and evaporation from soil and foliage
surfaces) for the vegetation (Figure 4b) against rainfall received indicated that
60% of the annual evaporation took place during a six month dry period when
virtually no rain was recorded. Since evaporation losses from dry soil and
shallow-rooted understorey species were minimal at this time, most of the
evaporative loss would have comprised transpiration of ground water by
actively growing phreatophytic tree and shrub species. Cumulative evapora-
tion loss for this six-month dry period amounted to approximately 400 mm,
or the equivalent of 47% of the total rainfall for the year. Conversely, total
evaporation for the other wet and cool six months of the year was only 270
mm and, with 85% of the year’s rain falling at this time, substantial recharge
of soil and water table would be expected, as was indeed observed during
this period.
In a somewhat similar study following mean daily rates of transpiration of
Banksia attenuata and B. menziesii at Melaleuca Park, 35 km north of Perth,
Dodd and Bell (1993) showed that tree water use was greatest in late summer
(rates up to 1.9 mm day–1 for B. attenuata and 0.3 mm day–1 for B. men-
ziesii), compared to only 0.3 mm day–1 and 0.05 mm day–1 for the same species
in winter. The bulk of the yearly transpiration from the tree storey again
occurred during months of little or no rain, a finding reinforcing the efficacy
of phreatophytic woody species in drying out soil profiles during summer
and autumn and in progressively lowering water tables back to pre-rain levels
of the autumn of the previous year. Data illustrating these principles (Dodd
and Bell, 1993) are detailed for B. attenuata in Figure 4(c) and (d), where it
can be seen how closely tree water use follows total global radiation over the
two years of study (1981 and 1982), and the marked extent to which tree
transpiration is out of phase with rainfall. Changes in levels of the water
table recorded at their site again reflect deep drainage in late winter and util-
isation of this water by phreatophytic trees the following summer.
From the limited data sets detailed above, it seems likely that rain
considerably in excess of transpirational demand will be received by Banksia
woodlands during an average winter season, and that this will promote deep
drainage to ground water (see also the article on similar phenomena in other
natural and agricultural systems by Hatton and Nulsen, this volume). Deep-
rooted components of the woodlands are thus of vital importance in drawing
on the winter surplus of water, firstly by drying out the soil profile, and then
by increasingly using ground water as the summer progresses. Where such
 320
Figure 4. (a) Monthly rainfall totals and (b) rates of evaporation for total vegetation and the three main vegetation components for Banksia woodland
near Gnangara, Western Australia during 1985 and 1986. (c) Depth to water table (line) and monthly rainfall totals (bars) and (d) mean daily values of
transpiration in Banksia attenuata (line) and total global radiation (TGR, bars) are also given for each month of the study period for Banksia woodland
of Melaluca Park, Western Australia during 1980 to 1982. Source: (a) and (b) redrawn from Farrington et al. (1989), (c) and (d) redrawn from Dodd
and Bell (1993).
 321

species are removed in land clearing and replaced by shallow-rooted annual

species, water tables inevitably rise dramatically across a landscape especially
in areas of low relief. For example, ground water levels under cleared land
at Moora are estimated to be increasing, over long term, at an average extent
of 10–20 cm per year (R. Speed, pers. comm.).

2.5. Trophic specialisations

Banksia woodlands occupy highly oligotrophic soils limited in readily avail-

able key elements, especially P and N. Tight nutrient cycling is evident from
the high efficiency of major and minor components of the flora in uptake
and usage of potentially limiting nutrients during growth, and in retrieval of
the same key elements prior to senescence of foliage (Lamont, 1984; Pate and
Dell, 1984; Pate, 1994). Such effectiveness leads to extremely low rates of
leaching of essential nutrients to ground water. At the Moora site, for instance,
nitrate and phosphate are barely detectable in ground water under pristine
woodland and seasonal peaks in mineralised N in the top 10 cm of soil of the
woodlands are very low, viz. 0.4 mg NO3––N and 0.6 mg NH4+–N kg–1 of soil
(Pate and Unkovich, in press). Not surprisingly, virtually all major compo-
nents of the flora show nutrient acquisition strategies of one form or another
in respect of P and N and probably also for other sparingly available nutri-
ents.
Possible trophic specialisations fostering effective nutrient uptake relating
generally to south western Australian heathlands and open woodlands have
been listed by Pate (1994). These include a variety of mycorrhizal and non-
mycorrhizal associations, symbiotic nitrogen fixation, and the ability to effect
heterotrophic inputs of organic and inorganic solutes through development
of carnivorous or parasitic habits. Families and species of plants subscribing
to these contrasting trophic categories are listed in Table 2 for the woodland
at the Moora and Kings Park sites. Virtually all nutrient acquisition strate-
gies are represented at one or both of the sites, but particularly common are
vesicular-arbuscular (VA) mycorrhizal woody shrubs (20% at Kings Park and
23% at Moora), woody species with non-mycorrhizal proteoid roots (16% at
Kings Park and 33% at Moora), VA- and orchid-type mycorrhizal herbaceous
species (16% at Kings Park and 11% at Moora) and various non-mycorrhizal
ephemeral and herbaceous taxa (23% at Kings Park and 14% at Moora).
Eighteen putative N2-fixing species were present at Kings Park and six such
species at Moora.
It is highly probable that competition for nutrients is intense between
species possessing the same acquisition strategies, especially in view of the
extremely low nutrient levels in the rooting substrate and the short season
during which soils are sufficiently moist for various uptake systems to operate
effectively. On the other hand, complementarity between different trophic cat-
egories is equally likely, especially where a critically important nutrient is
present in several different forms in the soil. Mutual benefits in terms of
 322
Table 2. Classification of trophic types encountered in Banksia woodlands examined at Moora and Kings Park amongst herbaceous (A) and woody (B)
species.

Moora Kings Park

(A) HERBACEOUS HABIT

Autotrophy Non-mycorrhizal e.g. Conostylis, Lomandra, e.g. Anigozanthos, Haemodorum
VA mycorrhizal Pityrodia (18 species) (24 species)
Orchid-type mycorrhizal Neurachne (1 species) Helichrysum, Stylidium (3 species)
Calaldenia (1 species) e.g. Diuris, Lyperanthus (8 species)
Nitrogen-fixing VA mycorrhizal None present Hardenbergia, Kennedia (2 species)
Carnivory Non-mycorrhizal Drosera (2 species) Drosera (1 species)
Holoparasitism Non-mycorrhizal Cassytha, Orobanche (2 species) Cassytha, Orobanche (2 species)
Heterotrophic epiparasitism Orchid-type mycorrhiza None present None present
Mycoautotrophy Ectomycorrhizal None present Lobelia (2 species)

(B) WOODY HABIT

Autotrophy Non-mycorrhizal (proteoid roots) e.g. Banksia, Hakea, Xylomelum e.g. Adenanthos, Grevillea
(16 species) (13 species)
VA mycorrhizal e.g. Boronia, Hibbertia, Pimelea e.g. Hibbertia, Macrozamia, Scaevola
(10 species) (7 species)
Ectomycorrhizal e.g. Actinostrobus, Eucalyptus e.g. Eucalyptus, Hypocalymma
(4 species) (7 species)
Ericoid mycorrhizal e.g. Astroloma, Lysinema e.g. Conostephium, Leucopogon
(3 species) (5 species)
Nitrogen-fixing VA mycorrhizal Acacia, Gompholobium (4 species) Acacia (3 species)
Ecto + VA mycorrhizal Allocasuarina (1 species) e.g. Gompholobium, Hovea (5 species)
Cluster roots (non-mycorrhizal) Jacksonia (2 species) Davesia, Jacksonia (4 species)
Root hemiparasitism Non-mycorrhizal Nuytsia (1 species) Nuytsia (1 species)
Aerial hemiparasitism Non-mycorrhizal None present Amyema (1 species)
 323

efficiencies of nutrient cycling may also apply where root distributions of

different species are mostly non-overlapping down a rooting profile (see Bell,
1995 for examples of such situations between epacridaceous, myrtaceous and
proteaceous species). The case studies from Kings Park (Pate et al., 1993;
Stewart et al., 1993) provide a good example of complementary patterns of
utilisation in respect of N resources, in demonstrating apparently mutualistic
interactions of (a) non-mycorrhizal herbaceous species relying principally on
nitrate, (b) N2 fixers subsisting on atmospheric N and (c) mycorrhizal woody
species presumably reliant on highly intractable organic forms of N. The latter
trophic specialisation is well exemplified by ericoid-type mycorrhizal asso-
ciations of Epacridaceae (Bell and Pate, 1996), whereas non-mycorrhizal
nitrate users are especially prominent after fire, when mineralisation of N of
belowground biomass of species killed in the fire provides transient large
resources of N (see Pate and Unkovich, in press).

2.6. Estimating proportional amounts of total biomass of major

2.6. components of a Banksia woodland ecosystem

Using mechanically-aided excavations of 3 m diameter × 3 m deep soil cores,

root biomass was exhaustively collected from three to ten replicate plants of
all common woody and large herbaceous species at the study site at Moora.
By comparing dry weights of total roots recovered with aboveground biomass,
shoot:root dry weight ratios were computed for all major species (Table 3,
column 2). The last fire at the site occurred 35 years previously (March 1962)
so the woodland was rated to be at or near post-fire maximum biomass
accumulation when sampled in 1997–98. Growth ring counts in trunks of
major seeder tree and shrub species indicated that virtually all major recruit-
ment occurred immediately after that fire (A. M. Grigg, J. S. Pate and M. J.
Unkovich (unpublished data)).
At Moora mean shoot dry weights of these same species were then assessed
by random sampling of 30 individuals of each species, and using the avail-
able data on shoot:root ratios (Table 3, column 2), values for mean standing
above- and belowground biomass (DM) per plant were computed for each
species (Table 3, columns 3 and 4). Density of each species per square metre
was assessed by random quadrat sampling throughout the bushland (Table 3,
column 5), and these values combined with previous biomass determinations
to estimate total standing biomass of each species per hectare (Table 3, column
6). The resulting data reinforce the prominence of biomass as large trees and
shrubs, and highlight the major impact that this is likely to have on nutrient
and water relationships at the site. It should be noted that it was not feasible
to recover biomass of seasonally-produced fine feeding roots of large woody
species, so estimates of belowground biomass of these are likely to be highly
conservative. This issue is referred to later when dealing specifically with
Banksia prionotes.
The total biomass of all species in the undisturbed woodland turned out to
 324
Table 3. Estimates of above and belowground biomass of 45 common woody species and the herbaceous perennial layer found in Banksia woodland
at Moora, Western Australia.

Species Shoot:root dry Biomass (kg ha–1) Density

weight ratio (no. plants ha–1)
Aboveground Belowground Total

Acacia pulchella 5.1 7 1 9 60

Actinostrobus arenarius 9.0 3290 366 3660 2510
Adenanthos cygnorum 4.1 1290 316 1610 461
Allocasuarina humilis 1.1 238 217 455 165
Astroloma xerophyllum 4.7 1 1 1 30
Banksia attenuata 1.5 454 309 763 185
Banksia leptophylla 4.0 685 171 857 296
Banksia prionotes 4.1 13100 2610 17800 1450
Beaufortia elegans 3.3 72 22 94 30
Boronia gracilipes 9.0 1 1 1 50
Calytrix sp. 1.2 23 19 42 687
Calytrix fraserii 2.5 34 14 48 687
Conospermum stochaedis 1.0 1 1 1 10
Cryptandra pungens 2.7 1 1 1 110
Dryandra nivea 1.0 8 8 16 60
Eremaea beaufortioides 3.5 1230 354 1580 1500
Eucalyptus todtiana 1.0 100 100 200 5
Grevillea eriostachya 4.6 1 1 1 20
Hakea obliqua 4.0 20 5 25 20
Hakea platysperma 4.0 20 5 25 20
Hakea trifurcata 3.8 23 6 29 10
Hibbertia acerosa 1.5 7 4 11 100
Hibbertia pachyrrhiza 2.8 2 1 3 360
Isopogon dubius 1.0 6 6 12 30
Isopogon scabriuscula 4.5 155 34 190 200
Jacksonia aff stricta 2.5 4 2 6 20
Jacksonia ulicifolia 1.8 63 36 100 190
Leptospermum erubescens 1.0 125 125 250 950
Leptospermum spinecens 0.7 8 13 21 30
Leucopogon polymorphus 4.3 14 3 17 650
Melaleuca scabra 1.5 1080 720 1800 2740
Melaleuca sp. 3.5 14 4 18 40
Mirbelia spinosa 4.1 7 2 10 170
Nuytsia floribunda 1.0 100 100 200 10
Petrophile brevifolia 0.6 15 26 42 50
Pileanthus sp. 3.5 65 19 83 687
Pimelea sp. 3.3 1 1 1 10
Pimelea spectabilis 1.6 1 1 1 10
Pityrodia bartlingii 1.3 56 42 97 680
Stirlingia latifolia 1.0 10 10 19 30
Synaphea spinosa 3.6 3 1 4 10
Verrauxia reinwardtii 1.0 1 1 2 10
Verticordia sp. A 3.2 2 1 3 91
Verticordia sp. B 4.4 5 1 7 91
Xylomelum angustifolium 1.7 306 182 489 120
Herbaceous layer1 (1.0) 995 995 1990 n.a.2
Mean ± SE 2.9 ± 0.0
Total (t ha–1) 24.5 7.8 34.5
1
Herbaceous biomass estimated as total dry weight of fully excavated plants per square meter. Data for aboveground biomass of herbaceous layer
estimated on basis of its shoot:root dry weight ratio being 1.0, in view of the extremely high proportion of cryptophytes and resprouter species.
2
Not measured.

325
326

be 32.5 t DM ha–1 (Table 3), with proportions of this as trees, woody shrubs,
herbaceous perennials and winter ephemerals (as shown in Figure 5A). Viewed
in terms of trophic specialisations, about two-thirds of the biomass was
associated with proteoid root-forming species, about one-seventh was vari-
ously mycorrhizal and only a few percent comprised of potential nitrogen-
fixing species (Figure 5B). Analysis in respect of seasonality of growth
showed most of the standing biomass to be associated with a summer pattern
of shoot extension, the remainder being winter-active. Virtually all of the

Figure 5. Relative proportions of biomass of species from the Banksia woodland at Moora,
Western Australia in terms of (a) life form as trees, woody shrubs, herbaceous perennials and
winter ephemerals, and (b) trophic specialisations including proteiod roots, mycorrhizal asso-
ciations and nitrogen-fixers.
 327

summer-active species were phreatophytic trees or shrubs, whereas most of

winter-active species were shallow-rooted (maximum depth of penetration 1.5
m) and therefore unable to access ground water at 5–6 m depth at the sites
concerned.

3. Banksia prionotes – major player of the woodlands

In this section we summarise the adaptive features exhibited by Banksia pri-

onotes – the arboreal dominant at our woodland site at Moora. The aim is to
provide a profile of the cardinal features deemed most desirable when selecting
a deep-rooted mimic species for the bushland.
Banksia prionotes is a fire-sensitive species recruiting most prolifically
after fire when dense, self-thinning stands quickly dominate post-fire vege-
tation of regions where it is a major arboreal component. Further recruitment
after this post-fire germination response is generally insignificant. For
example, counts of trunk growth rings of B. prionotes in the woodland at
Moora indicated that virtually all trees currently present had recruited from
a hot summer fire in the area 35 years previously. According to studies of
populations at Yanchep, growth of individuals in post-fire stands of B. pri-
onotes increases exponentially over the first eight or so years after germina-
tion, with shoot:root dry weight ratios increasing from two to three in young
plants to a plateau value of 3.8–4.5 for plants of four years of age or older
(Bowen and Pate, 1991; Pate et al., 1998). Distribution of dry matter between
plant parts shows fastest relative growth rates to be accompanied by peak
ratios of leaf area:total plant weight in plants of three to four years of age
and greatest rates of accumulation of dry matter per unit leaf area at the same
time.
The dimorphic rooting character of B. prionotes, alluded to in this article
and that of Pate and Dawson (this volume) becomes evident early after
germination and tap (sinker) roots of surviving seedlings would typically have
already penetrated to 2 m depth by the end of their first summer. Tap roots
of the species eventually descend to 15 m or more if a more elevated water
table is absent, or upon meeting the capillary fringe of a water table, will
branch out in a complex network of fine water-absorbing roots. Seedlings
show an initial complement of three to five lateral roots radiating from the
top of the root stock and before a tree reaches ten years of age, eight to ten
lateral roots should have developed. Mapping of locations and branching
patterns of these roots (Pate et al., 1998) reveals effective exploitation of
space, to the extent that rooting catchments of mature trees will generally
occupy ground areas four to six times greater than that shaded by their crowns.
Roots of older trees at typical woodland densities are thus likely to trespass
to a great extent into rooting zones of neighbours. Competition for water and
nutrients is therefore viewed as a major contributing factor to the progres-
sive thinning of stands and slowing down of growth rates in middle-aged
328

populations of the species. Most trees at Moora show ‘secondary’ tap roots
descending from mid regions of one or more of their major lateral roots
(J. S. Pate and S. S. O. Burgess, unpubl. data). These auxilliary tap roots even-
tually access ground water just as effectively as does the primary tap root.
Seasonal production of proteoid roots commences each autumn following
outgrowth of new initials from periderm-invested perennating parts of the
lateral root system. There is evidence from deuterium:hydrogen (D/H) studies
and heat-pulse experiments (Dawson and Pate, 1996; Pate and Dawson, this
volume) that water hydraulically lifted by tap roots supports early growth of
feeding roots in autumn, maintains integrity of proteoid root clusters through
transient drought in winter and spring, and extends activity of such roots
thereafter well into a summer season. This facility would clearly ameliorate
the vulnerability of superficially (5–15 cm) located feeding roots to sudden
change in soil water status. It is not known whether shallow-rooted species
benefit appreciably from hydraulically-lifted water.
Study of seasonal changes in xylem sap composition of lateral roots, sinker
roots, trunks and proteoid root clusters of B. prionotes, taken together with
corresponding analyses of seasonal filling and emptying of storage pools of
minerals in trunks and leaves, have provided an overall picture of mineral
resource partitioning (Pate and Jeschke, 1993). Throughout their peak activity
in late winter and spring proteoid roots comprise principal contributors of
PO43–, amino acids and K+ to the xylem stream and may also function
significantly in reducing nitrate released following sporadic nitrification events
in the upper soil profile (Jesckhe and Pate, 1995). Xylem sap of laterals in
receipt of minerals from proteoid roots is therefore consistently more con-
centrated in a range of essential nutrients, such as N and P, during the wet
season than is that of the tap root. Xylem sap collected simultaneously from
the trunk base shows mineral concentrations intermediate between those of
lateral and tap roots, implying that the shoot concerned is being supplied
with a mixture of both xylem streams, more or less as predicted from the
D/H studies detailed by Dawson and Pate (1996) Pate and Dawson (this
volume). Analyses of gradients in xylem sap concentrations of P, N and other
nutrients from base to top of trunks of trees harvested throughout a year
provide evidence of lower parts of stems withdrawing nutrients laterally from
the xylem during winter and accommodating these nutrients in storage
parenchyma of xylem, especially that of rays (Pate and Jeschke, 1995). The
following summer through to autumn, nutrients are released back to the xylem,
and together with nutrients mobilised from oldest leaves, provide a major
resource base for nutrition of the rapidly extending shoot. Further uptake of
minerals from the root is of course limited during summer, due to (a) loss of
proteoid roots with drying out of the upper soil, and (b) general deficiencies
of macronutrients in the ground water source accessed by the tap root.
However, large quantities of water from deep sources remain vital for
maintenance of plant water status and, more importantly, for enabling high
rates of photosynthesis to implement shoot growth in summer. These features
 329

remain contingent upon tap roots drawing successfully from the retreating
water table. Indeed, where this does not happen, sudden death of trees is
observed, especially in late autumn following a previous winter of poor
rainfall. Presumably, the more vigorous surviving trees had lowered ground
water to levels beyond the reach of inferiors which have succumbed.
The complex hydraulic architecture of xylem of trunk and roots of B.
prionotes and other dimorphic-rooted proteaceous species (Pate et al., 1995)
is viewed as an essential ingredient for maintenance of integrity of vessel
functioning through seasonal conditions in which large scale changes occur
in water status of upper soil and atmospheric water deficit. The very long
and wide vessels of the tap root, together with increases in the specific
hydraulic conductivity of xylem with downward attenuation in root diameter,
result in the xylem pathway from ground water to soil level being of uniformly
very low resistance. In marked contrast, the much shorter and narrower vessels
of the trunk give high resistances to flow, thus comprising a large buffer
resistance against which stomata can respond to transient or long term water
stresses. It remains to be seen whether some of the exotic species of trees
which we suggest later as possible mimics of Banksia woodlands possess the
same adaptive qualities.
Particularly relevant to replacement of natural vegetation with a functional
equivalent mimic is the requirement to evaluate the net primary productivity
of biomass of the original bushland and relate this to annually available
resources of rainfall and nutrients. We have already shown that standing living
above- and belowground biomass of the Moora woodland was rated at approx-
imately 32.5 t DM ha–1 and that 55% (17.8 t DM ha–1) of this consisted of
the single dominant B. prionotes, and a further 16% (5.1 t DM ha –1) com-
prised of other woody trees or shrubs of 2 m or more in height. Our protocol
for estimating net annual primary production of these woody elements was
as follows.
Due to a strictly summer pattern of growth in which seasonal flushes of
new shoots and foliage persist on the plant for three seasons, this last three
years worth of extension growth to main and side branches of tree or shrub
species of genera such as Banksia, Xylomelum, Adenanthos or Hakea can be
readily identified and collected separately from older biomass. Using this
protocol the last three years of shoot growth of B. prionotes at Moora was
estimated as equivalent to 21.1% (or 2.7 t DM ha–1) of the current above-
ground biomass of the trees (see Figure 6). By cutting out a segment from
the trunk base at 10 cm or so aboveground level and examining its annual
growth rings, the last three years of secondary growth can be easily chiselled
off from a circumferential sector of the outer trunk and the dry weight of this
off cut compared with that of the remaining earlier formed wood of the sector.
In this way net productivity in terms of secondary thickening was assessed
for B. prionotes as 8.8% (or 1.0 t DM ha–1) of the wood biomass currently
present in the trunk (Figure 6). Then, assuming that secondary thickening of
the perennial superstructure of the root increased similarly (see Figure 6 roots),
330

Figure 6. Model depicting dry matter in total standing biomass and biomass components relating
to three years leafy growth, annual proteoid root production and secondary thickening of Banksia
prionotes (1995–1997) in pristine bush at Moora, Western Australia. Annual net primary
production (ANPP) was determined on the basis of total current standing biomass including
assumed proteoid root production averaged for the three year period of the study.

our earlier estimate of 17.8 t DM ha–1 for B. prionotes at the site would have
effected a net biomass increment of 3.9 t DM over the past three years, made
up of 2.7 t of new shoot extensions and 1.2 t DM of growth increase in trunk
and perennial parts of the root system. We then need to add to this a further
estimate of the new biomass laid down each year in annual investment of
fine feeding roots, predominantly composed of proteoid root biomass
(Figure 6). As suggested from previous evaluations (Pate et al., 1998) we
tentatively assess this as representing the equivalent of a total investment of
46% of the current biomass of perennial parts of the root system over the three
year period. Rated as such, it is given a value of 1.2 t DM ha–1 for the three
 331

year study period depicted in Figure 6. Using the data represented pictorially
in Figure 6, our model accordingly suggests a net above- plus belowground
productivity of 1.7 t DM ha–1 year–1 for the population of B. prionotes at the
site, and a yearly net increment equivalent to 10.6% of the present standing
biomass.
The above approach (Figure 6), involving a three year rather than a single
year estimate of net primary production of biomass, carried the advantage of
averaging tree response at Moora over a first year of well above average (1995
– 703 mm) rainfall, a second one of near average (1996 – 448 mm) rainfall
and a third (1997 – 431 mm) of marginally less, most of which fell in summer
at times of high evaporation from the soil surface. Leaf weights on each
summer’s annual shoot extension of this chronosequence varied markedly, viz.
in a weight ratio of 11 (1995):7 (1996):3 (1997), i.e. more or less reflecting
the effectiveness of rainfall of the previous winter in replenishing water in
soil and ground water and mobilising soil nutrients from soil to standing
biomass.
Our approach to measuring net primary productivity proved to be applic-
able to any summer-active woody species that showed tell-tale morpholog-
ical evidence of seasonal growth increments. It was encouraging in this
connection to find similar proportional increases in shoot biomass over the
three year period to that recorded above when this was applied to other species
at the Moora site (viz. 24% for B. leptophylla, 26% for B. attenuata and 24%
for Adenanthos cygnorum versus the already mentioned value of 21% for B.
prionotes).
Assuming the above approach to apply generally to all biomass at the site
we estimated total net primary productivity for the Banksia woodland system
as 3.2 t DM ha–1 y–1 (i.e. 10.6% of the 32.5 t DM ha–1 of total biomass shown
in Table 3). With a mean annual rainfall of 527 mm for the three years studied,
net productivity of the ecosystem then translates to 6.1 kg of total biomass
ha–1 mm–1 of annual rainfall, or 4.5 kg DM ha–1 mm–1 if estimated solely in
terms of aboveground biomass.
Matching data on the productivity of total aboveground biomass for pasture,
lupin and wheat for the period 1994–96 at a location a few kilometres from
the Banksia woodland study site (see data of Anderson et al., 1998a and the
unpublished observations of E. C. Lefroy and J. S. Pate for a tagasaste trial
adjacent to the Moora bushland site) are given alongside those for the Banksia
woodland in Table 4. A somewhat surprising feature of the comparison is
that annual productivity per unit rainfall of the bush (4.5 kg DM mm–1) is least
of all values recorded, with the exception of one particularly poor wheat crop
(4.0 kg DM mm–1, 1995 season). The border of the pristine bush at the study
site receives run-off from nutrient-rich water of the elevated water table on
adjacent agricultural land, and trees of B. prionotes and other nutrient-tolerant
native species in this ‘hedge’ of vegetation have achieved gigantic propor-
tions embodying productivity rates of up to 15.9 kg DM mm–1 rainfall (see
Table 4). However, even in this cossetted situation of much freer access to
332

Table 4. Net annual primary productivity of aboveground biomass of annual pasture and crops,
a plantation of tagasaste (Chamaecytisus proliferus) and neighbouring Banksia woodland in the
Moora region of Western Australia.

Net annual productivity Biomass production

of aboveground biomass per unit rainfall
(t ha–1) (kg mm–1)

1994: 294 mm annual rainfall

Lupina 5.3 18.0
Wheata 3.5 11.9
1995: 703 mm annual rainfall
Lupina 6.5 9.2
Wheata 3.7, 2.8 5.3, 4.0
Pasturea 7.3, 7.6 10.3, 10.8
1996: 438 mm annual rainfall
Lupina 5.8 13.2
Lupinb 4.2 9.6
Wheata 7.5, 6.8 17.1, 15.5
Pasturea 4.2, 5.0 9.6, 11.4
Pastureb 3.4 7.8
Tagasasteb, e 4.4 10.0
1997: 412 mm annual rainfall
Oatsb 3.1 7.5
Tagasasteb 20.1 48.7
1995 – 1997: 527 mm mean annual rainfall
Banksia woodlandc 2.4 4.5
Hedge of Banksia woodland bordering
agricultural landd 8.4 15.9

Source and notes: a Data from Anderson et al. (1998a); b Data from E. C. Lefroy and J. S. Pate
(unpubl. data); c Data derived from pristine bush unaffected by agricultural runoff as shown in
text and Table 3; d Receives subterranean runoff of nitrate-polluted water throughout the year
from elevated water table in agricultural land (A. Grigg, D. J. Arthur, T. E. Dawson, T. J. Hatton,
M. J. Unkovich and J. S. Pate, unpubl. data); e Tagasaste cut in May 1996 and not cut in 1997.

water and nutrients, the Banksia and other trees still fail to out-produce the
best wheat crops shown in the data set, and achieve a net annual production
of aboveground biomass equivalent to only 40% of that of a tagasaste
(Chamaecytisus proliferus Link ssp. palmensis) plantation in its second year
after cutting (Table 4; see also Lefroy and Stirtzaker, this volume).
The apparently greater productivity of the annual crops than Banksia in
terms of rainfall received becomes even more surprising when one considers
that deep drainage losses of unused water occur during growth of the former
(see data for deep sands of the Moora region by Anderson et al., 1998b;
Asseng et al., 1998), whereas, seasonal rises and falls in water table in the
woodland area indicate long-term balance of rainfall utilisation and therefore
a more meaningful relationship between net primary productivity and rainfall.
We conclude that this alleged ‘ineffectiveness’ of the native bush in dry matter
 333

production relates to the perennial character of the vegetation and the major
constraints on growth likely to be imposed by the highly oligotrophic nature
of the rooting substrate. In relation to perenniality, the model of Figure 5
shows relatively large annual investments of dry matter in secondary thick-
ening of root and trunk and feeding roots essential for acquisition of
nutrients in the wet season. We have little idea of respiratory costs in
construction, maintenance and functioning of these heterotrophic parts of the
plant, but judging from budgets of allocation of photosynthate for proteaceous
seedlings including B. prionotes (see Bowen and Pate, 1991), respiration losses
of carbon are at least equal to that allocated to dry matter. It is reasonable to
assume that maintenance costs would be much less in annual crop plants, given
the ephemeral character of their shoots and roots and the much higher
proportions of their biomass as photosynthesising leaves as opposed to stems
and roots in comparison to life forms such as large and old woody trees.
Furthermore, costs of nutrient acquisition are likely to be much less in fertile
agricultural land due to fertiliser inputs and greater availability of recycled
nutrients. Also, where winter annual crops are deployed there is no require-
ment for deep long-lived roots to be developed to combat summer heat and
drought and access ground water.
Notwithstanding the above, it is still difficult to explain the marked dis-
crepancy in productivity between the ‘hedge’ of Banksia receiving nutrient
run off from the surrounding cleared land, and performance in the second year
after cutting of tagasaste plantations which benefited similarly as sumps for
water and nutrients from a hinterland of pasture and annually cropped areas.
Cutting of the tagasaste the previous year, and the relative youth of the trees
concerned relative to the near climax community of Banksia trees of the bush,
might be identified as the critical ingredient for much higher productivity of
the former than latter system. Thus, as shown for a number of native woody
species of similar deep-sand habitats (e.g. Hansen et al., 1991; Bowen and
Pate, 1991, 1993; Bell and Pate, 1996), destruction of shoots by fire or cutting
can elicit extremely fast rates of shoot growth as shoot:root biomass ratios are
restored to pre-disturbance levels. With relatively small stem biomass and high
ratios of leaf mass to total plant mass at such times, maintenance costs should
be relatively low leading to high effectiveness in dry matter production, even
possibly equalling that of juvenile plants. We would suggest that the cut
plantation of tagasaste was in such a highly productive, yet water expensive,
state when sampled in the 1997 season at Moora.

4. Towards a mimic system for Banksia woodlands

4.1. Radical changes following clearing for agriculture

Banksia woodlands, or in more general terms, Banksia sand ecosystems of

Western Australia were initially avoided by farmers in view of their highly
334

oligotrophic character, and did not become amenable to economically-viable

cropping or pasture production until specific trace element additions together
with superphosphate were applied to boost yields. Such areas are now mostly
devoted to N2 fixing pasture and crop legumes in rotation with cereals. As
alluded to earlier, annual crops on these deep sands fail to utilise seasonal
rainfall, leading to particularly severe examples of rising ground water levels.
Although not usually affected by salinity, cleared deep sands also provide
examples of alarmingly high levels of nitrate pollution of ground water. For
example, Pate and Unkovich (in press) in their comparison of N2 fixing inputs
in the natural bushland and adjacent ecosystems at Moora, found groundwater
under pasture or cereals to contain 6–12 mg NO3––N L–1 versus only 0.2 mg
NO3––N L–1 under neighbouring undisturbed native bush. This nitrate in
cleared land has clearly been generated by agricultural practices and leached
progressively to ground water since first clearing of the bush in the region in
the late 1950’s and 1960’s. Furthermore, since nitrogen fertilisers are used
only moderately for cereal production in the region, the N pollution currently
present is mostly attributable to mineralisation of the easily decomposed
biomass residues of N2 fixing leguminous crops and pasture components. With
virtually zero anion-retaining capacity in the rooting substrate, and significant
episodes of a season’s mineralisation occurring out of phase with principal
periods of crop and pasture growth, nitrate leaching may be substantial and
more pronounced under certain crops such as lupins than under pastures
(Anderson et al., 1998b). In any event, with additions of P and trace elements,
unbridled inputs of fixed N, elevated water tables and nitrate leaching, the
agricultural scenario features nutritional and hydrological states which are
grossly perverted from the balanced and self-sustaining natural ecosystems
originally present in the area. Herein lies the crux of the problem when
designing ‘alternative’ farming systems suited for rehabilitation of such areas.

4.2. Two contrasting scenarios for application of mimic principles

The above issues raise the question of what systems one should address when
developing the mimic principal. One strictly limited scenario would be, say,
where one wishes to revegetate a Banksia sand ecosystem which had been
essentially denuded of biomass in timber harvesting, sand mining or some
form of clearing, but where no fertiliser inputs had been involved and no
eutrophication incurred through run-off from adjacent areas. Under such
circumstances, there would surely be no better solution than to replant with
a selection of major indigenous species, and to do so with judicious consid-
eration of proper balance between life forms, rooting morphologies, phe-
nologies of growth and trophic specialisations present at the site prior to
disturbance.
Judging from the universally high prevalence of deep-rooted phreatophytic
shrubs and trees in Banksia sand ecosystems, our first priority would be to
select these as principal players. Planting would then be designed to achieve
 335

effective balance between summer-active deep-rooted versus winter-active

shallow-rooted types, and hence, maintain an overall balance in seasonal water
utilisation. It remains to be seen how critical the planting ratio between these
two types is, or whether, through thinning and attrition, the same balance
would result more or less regardless of proportional representations at the time
of planting. It would also be of value to follow short- and long-term impli-
cations on water table levels and nutrient balances in systems where bias was
artificially maintained towards shallow-rooted winter-active species over
phreatophytic summer-growing species, and vice versa. In essence, we suggest
two perennial functional types; summer-active deep-rooted and winter-active
shallow-rooted (see Character option 1, Figure 7), with winter-active deep-

Figure 7. A four stage species selection suggested for a mimic system to replace vegetation of
cleared Banksia woodlands in south western Australia. Numbers of species found in Moora
(M) and Kings Park (KP) fitting the particular character states are indicated. Fire response is
indicated as resprouter (R) or seeder (s).
336

rooted and summer-active shallow-rooted species essentially not represented

in the selection, or indeed to any extent in the original vegetation. Numbers
of species available in the natural ecosystems to cover these two types are
shown in Figure 7.
A next consideration (Character option 2, Figure 7) would be to include
under each of the above character states, non-woody herbaceous species and
small shrubs along with trees to provide a more or less continuous ground
cover. Most Banksia woodlands have good representation of such smaller
species. However, it turns out that the floras of Moora and Kings Park do not
include species which are deep-rooted summer-active perennial herbs or
shallow-rooted winter-active trees. These types, indicated by dotted lines in
Figure 7, are accordingly excluded from further consideration.
Thirdly, the menu for species selection proposed incorporates resilience to
fire by balanced representation of fire-sensitive (seeder) and fire-resistant
(resprouter) species (Character option 3, Figure 7). Large trees and shrubs
come in both categories and preponderance of one over the other varies with
site, as shown in the comparison between Moora and Kings Park floras in
this article. However, herbaceous perennial species are dominantly resprouters
and their value in holding topsoil together and their activity, and that of
herbaceous ephemerals, in reabsorbing minerals released after fire cannot be
overestimated. Slow growth of resprouters as seedlings, scarcity of seed stocks
and poor capacity to compete initially with faster growing seeders, can cause
problems when trying to achieve adequate stocking of fire-resistant species
on cleared land. Again, experimentation within this area is clearly called for.
At this stage our choice portfolio (Figure 7) contains eight sets of variously
combined character states and we would regard this as possibly a basic
minimum to successful functioning of the replacement system. As can be see
from Figure 7, the floras of Moora and Kings Park contain at least one or
more representations in each of the eight categories and each could, therefore,
be used feasibly in revegetation programs.
Finally, we would envisage a further sophisticated series of possible options
to amplify species selection and that these should be based on major trophic
specialisations within the original floras (i.e. proteoid roots, various mycor-
rhizal types, symbiotic N2 fixation, carnivorous and parasitic habit). With
this further set of options (Table 5), the species listing escalates to 48 possible
options, a number of which are minimally represented or not represented at
all in the floras of the study sites.
There are two obvious dangers in the above approach. Firstly, it would be
unwise to include only one representative of each set of character states,
simply because it might either fail to establish to the desired density or because
it was a poor choice from among the replicate candidates available carrying
that set of options. Secondly, the scheme presented assumes that the selected
sets of character options are the only ones necessary for the mimic to function
as a fully sustainable equivalent to the indigenous system. In the absence of
information on levels of other possibly equally essential species interactions
Table 5. Possible options for species selection for replacement model based on trophic specialisations after previous selection of species from Character
options 1–3 outlined in Figure 7.

Trophic type Deep-rooted summer-active Shallow-rooted winter-active

Tree Shrub Shrub Perennial herb

S1 R1 S R S R S R

VA mycorrhiza M 1 3 3 3 No sp. No sp. No sp. No sp.

KP 1 4 0 3
Other mycorrhiza2 M No sp.3 0 1 3 7 1 3 3
KP 3 1 3 1 7 2 10
Proteoid roots M No sp. 1 No sp. 1 0 0 No sp. No sp.
KP 3 1 1 2
N2 fixation M No sp. 0 1 3 0 0 No sp. No sp.
KP 1 1 3 1 1
Carnivory M No sp. No sp. No sp. No sp. No sp. No sp. 1 1
KP 0 1
Parasitism M No sp. 1 No sp. No sp. 0 No sp. 1 No sp.
KP 1 1 2
1
S = seeder; R = resprouter.
2
Includes ectomycorrhiza, ericoid-type mycorrhiza, orchid mycorrhiza.
3
No sp. = no species were recorded for this choice.

337
338

in the bushland, the protocol of choice of players indicated in Figure 7 and

extended into Table 5, might well lead to dysfunctionality, especially if the
final biomass balance between physiognomic groupings turned out to be very
different to that of the original bush.
A second entirely different scenario concerns species selection and planting
designs directed at amelioration of the substantial hydrological and nutritional
balances ‘within an agricultural setting’ or ‘under agricultural use’ or ‘com-
mercial use’. We are hesitant to use the term ‘mimic’ in such a case, since
the desired objectives of the proposed system in the first instance would be
to make progressive inroads into elevated water tables and excessive nitrate
and other nutrient loadings in the system. This might well require use of
summer-active deep-rooted tree or shrub species which are much faster
growing than those of native bushlands, and species which would also prove
amenable to planting in alley or plantation formats, where they would sup-
posedly achieve greater utilisation of ground water and nutrients per unit
planted area than is currently experienced in annual crop species, or indeed
in native bush. A number of exotic tree species, such as Pinus pinaster, P.
radiata, Paulownia sp., various mallee eucalypts, Acacia saligna (for rela-
tively saline areas) and the use of tagasaste over the past decade or so come
to mind in this connection. Were row plantings of such species to be accom-
panied by regular cropping with winter-active cereals and legumes, as
currently practised with tagasaste, desirable outcomes might well be achieved
in both short and long term. However, two important caveats would apply.
Firstly, we would view with mistrust continued use of shrub legumes such as
tagasaste, since when these are cut or grazed annually for fodder they typi-
cally make a greater net contribution of fixed N2 to the system than they utilise
from soil and groundwater in the form of nitrate-nitrogen (see case studies of
Pate and Unkovich (in press) for tagasaste at Moora). Indeed, unless one were
to select a non-nodulating genotype of tagasaste, or to genetically engineer
ineffective nodulation of the species, its use, while clearly desirable in terms
of hydrological balance, might exacerbate rather than ameliorate current nitrate
pollution problems. Secondly, use of shallow- and deep-rooted species for
nutrient stripping in terms of nitrate and possibly also excess P, would be
accomplished only if the nutrients in question were removed from the site as,
say, foliage, grain, animal products, or similar. This is unlikely to eventuate
in the case of unpalatable shrubs and trees unless a commercial incentive were
to accrue in the form of harvestable timber, biomass of wood chipping, or
possible specific highly valuable products such as oil from leaves, as in the
case of mallee eucalypts. As long as government agencies fail to subsidise
environmental rehabilitation on a scale attractive to landowners, options avail-
able to farmers remain strictly limited, especially in current gloomy economic
times for the agricultural industry. However, current discussions on trading
of carbon credits might well ultimately provide extra profit incentive in this
connection. In any event, vast scales of operation, enormous investments,
and an unprecedented degree of cooperation between landowners will be
 339

required to plant whole catchments with sufficient deep-rooted woody species

to redress the crucial problems on the deep sand ecosystems of Western
Australia.
Finally, another option to be considered is that of utilising solely herba-
ceous species for rehabilitation programmes, in this case using annual winter
crops and pastures as currently employed, alongside novel plantings of deep-
rooted summer-active herbaceous perennials. Lucerne (Medicago sativa),
despite possible disadvantages in view of symbiotic inputs of N, and summer-
active C4-type grasses such as Rhodes grass (Chloris guyana) would be
obvious candidates for summer usage of water and nutrients, but their success
would obviously depend on capacity to access underlying ground water. Both
species root to at least 4–5 m in depth in deep sands, but their trialling is
still very limited in the highly depauperate acidic sands of Western Australia.
Despite obvious non-comparability in every respect against the sclerophyl-
lous woodlands which they would replace, such all-herbaceous mimics might
offer distinct advantages in terms of ease of management and profitability
relative to alley or plantation tree-based systems. Indeed, if water table rise
continues unabated in certain areas, relatively shallow-rooted herbaceous
perennials might offer the only viable option for dealing with near surface
ground water tables, where flooding-intolerant woody species would be
unlikely to survive.

5. Acknowledgements

The authors wish to acknowledge the use of a great deal of as yet, unpub-
lished data generously supplied by our various collaborators in the Moora
project: T. E. Dawson, T. J. Hatton, S. S. O. Burgess, E. C. Griffin, D. J. Arthur,
M. J. Unkovich, A. M. Grigg and E. C. Lefroy. Thanks are also due to A. J.
Smit for assistance in preparation of the figures. Funding of research by grants
from the Land and Water Resources Research and Development Corporation,
the Australian Research Council and the Department of Conservation and Land
Management, Western Australia is gratefully acknowledged.

6. References

Anderson GC, Fillery IRP, Dolling PJ and Asseng S (1998a) Nitrogen and water flows under
pasture-wheat and lupin-wheat rotations in deep sands in Western Australia. 1. Nitrogen
fixation in legumes, net N mineralisation, and utilisation of soil-derived nitrogen. Australian
Journal of Agricultural Research 49: 329–343
Anderson GC, Fillery IRP, Dunin FX, Dolling PJ and Asseng S (1998b) Nitrogen and water
flows under pasture-wheat and lupin-wheat rotations in deep sands in Western Australia. 2.
Drainage and nitrate leaching. Australian Journal of Agricultural Research 49: 345–361
Asseng S, Fillery IRP, Anderson GC, Dolling PJ, Dunin FX and Keating BA (1998) Use of the
APISM wheat model to predict yield, drainage, and NO3– leaching for a deep sand. Australian
Journal of Agricultural Research 49: 363–377
340

Beard JS (1984) Biogeography of the kwongan. In: Pate JS and Beard JS (eds) Kwongan, Plant
Life of the Sandplain, pp 1–26. University of Western Australia Press, Nedlands
Beard JS (1989) Definition and location of the Banksia woodlands. Journal of the Royal Society
of Western Australia 71: 85–86
Bell DT, Hopkins AJM and Pate JS (1984) Fire in the kwongan. In: Pate JS and Beard JS (eds)
Kwongan, Plant Life of the Sandplain, pp 178–204. University of Western Australia Press,
Nedlands
Bell TL (1995) Biology of Australian Epacridaceae: With Special Reference to Growth, Fire
Response and Mycorrhizal Nutrition. PhD Thesis, The University of Western Australia
Bell TL and Pate JS (1996) Nitrogen and phosphorus nutrition in mycorrhizal Epacridaceae of
south-west Australia. Annals of Botany 77: 389–397
Bell TL, Pate JS and Dixon KW (1994) Response of mycorrhizal seedlings of SW Australian
sandplain Epacridaceae to added nitrogen and phosphorus. Journal of Experimental Botany
45: 779–790
Bell TL, Pate JS and Dixon KW (1996) Relationships between fire response, morphology, root
anatomy and starch distribution in South-west Australian Epacridaceae. Annals of Botany
77: 357–364
Bettenay E (1984) Origin and nature of the sandplains. In: Pate JS and Beard JS (eds) Kwongan,
Plant Life of the Sandplain, pp 51–68. University of Western Australia Press, Nedlands
Bowen BJ and Pate JS (1991) Adaptations of SW Australian members of the Proteaceae:
allocation of resources during early growth. Proceedings of the International Protea
Association, Sixth Biennial Conference, Perth, Western Australia, pp 347–356. Promaco
Conventions Pty Ltd
Bowen BJ and Pate JS (1993) The significance of root starch in post-fire shoot recovery of the
resprouter Stirlingia latifolia R. Br. (Proteaceae). Annals of Botany 72: 7–16
Dawson TE and Pate JS (1996) Seasonal water uptake and movement in root systems of
Australian phreatophytic plants of dimorphic root morphology: a stable isotope investiga-
tion. Oecologia 107: 13–20
Dodd J and Bell DT (1993) Water relations of the canopy species in a Banksia woodland, Swan
Coastal Plain, Western Australia. Australian Journal of Ecology 18: 281–293
Dodd J and Griffin EA (1989) Floristics of the Banksia woodlands. Journal of the Royal Society
of Western Australia 71: 89–90
Dodd J and Heddle EM (1989) Water relations of Banksia woodlands. Journal of the Royal
Society of Western Australia 71: 91–92
Dodd J, Heddle EM, Pate JS and Dixon KW (1984) Rooting patterns of sandplain plants and
their functional significance. In: Pate JS and Beard JS (eds) Kwongan, Plant Life of the
Sandplain, pp 146–177. University of Western Australia Press, Nedlands
Farrington P, Greenwood EAN, Bartle GA, Beresford JD and Watson GD (1989) Evaporation
from Banksia woodland on a groundwater mound. Journal of Hydrology 105: 173–186
Farrington P and Bartle GA (1991) Recharge beneath a Banksia woodland and a Pinus pinaster
plantation on coastal deep sands in south Western Australia. Forest Ecology and Management
40: 101–118
Gardner CA (1944) The vegetation of Western Australia with special reference to the climate
and soils. Journal of the Royal Society of Western Australia 28: 11–87
Gozzard JR and Mouritz MJ (1989) Mineral resources and mining of the Spearwood and
Bassendean Dune Systems. Journal of the Royal Society of Western Australia 71: 109–110
Hansen A, Pate JS and Hansen AP (1991) Growth and reproductive performance of a seeder
and a resprouter species of Bossiaea as a function of plant age after fire. Annals of Botany
67: 497–509
Hobbs RJ and Atkins L (1990) Fire-related dynamics of a Banksia woodland in south-western
Western Australia. Australian Journal of Botany 38: 97–110
Hopkins ER (1989) Forestry and Banksia woodlands on the Swan Coastal Plain. Journal of the
Royal Society of Western Australia 71: 107–108
Jeschke WD and Pate JS (1995) Mineral nutrition and transport in xylem and phloem of Banksia
 341

prionotes and proteaceous trees of dimorphic root morphology. Journal of Experimental

Botany 46: 895–905
Lamont BB (1984) Specialised modes of nutrition. In: Pate JS and Beard JS (eds) Kwongan,
Plant Life of the Sandplain, pp 236–245. University of Western Australia Press, Nedlands
Pate JS (1994) The mycorrhizal association: just one of many nutrient acquiring specialisations
in natural ecosystems. In: Robson AD, Abbott LK and Malajczuk N (eds) Management of
Mycorrhizas in Agriculture, Horticulture and Forestry, pp 1–10. Kluwer Academic Press,
Netherlands
Pate JS, Casson NE, Rullo JC and Kuo J (1985) Biology of fire ephemerals of the sandplains
of the kwongan of south-western Australia. Australian Journal of Plant Physiology 12:
641–655
Pate JS and Dell B (1984) Economy of mineral nutrients in sandplain species. In: Pate JS and
Beard JS (eds) Kwongan, Plant Life of the Sandplain, pp 236–245. University of Western
Australia Press, Nedlands
Pate JS, Dixon KW and Orshan G (1984) Growth and life form characteristics of kwongan
species. In: Pate JS and Beard JS (eds) Kwongan, Plant Life of the Sandplain, pp 84–100.
University of Western Australia Press, Nedlands
Pate JS, Froend RH, Bowen BJ, Hansen A and Kuo J (1990) Seedling growth and storage
characteristics of seeder and resprouter species of Mediterranean-type ecosytems of S.W.
Australia. Annals of Botany 65: 585–601
Pate JS and Jeschke WD (1993) Mineral uptake and transport in xylem and phloem of the
proteaceous tree Banksia prionotes. Plant and Soil 155: 273–276
Pate JS and Jeschke WD (1995) Role of stems in transport, storage and circulation of ions and
metabolites by the whole plant. In: Gartner B (ed) Stems and Trunks: Their Roles in Plant
Form and Function, pp 177–204. Academic Press, New York
Pate JS, Jeschke WD and Aylward MJ (1995) Hydraulic architecture and xylem structure of
the dimorphic root systems of South-West Australian species of Proteaceae. Journal of
Experimental Botany 46: 907–915
Pate JS, Jeschke WD, Dawson TE, Raphael C, Hartung W and Bowen BJ (1998) Growth and
seasonal utilisation of water and nutrients by Banksia prionotes. Australian Journal of Botany
46: 511–532
Pate JS, Raisins E, Rullo J and Kuo J (1986) Seed nutrient reserves of Proteaceae with special
reference to protein bodies and their inclusions. Annals of Botany 57: 747–770
Pate JS, Stewart GR and Unkovich MJ (1993) 15N natural abundance of plant and soil
components of a Banksia woodland ecosystem in relation to nitrate utilization, life form,
mycorrhizal status and N2-fixing abilities of component species. Plant, Cell and Environment
16: 365–373
Pate JS and Unkovich MJ (in press) Measuring symbiotic nitrogen fixation – Case studies of
natural and agricultural ecosystems in a Western Australian setting. In: Advances in
Physiological Plant Ecology, Proceedings of the 1998 British Ecological Society Symposium,
York, September 1998
Semeniuk V and Glassford DK (1989) Bassendean and Spearwood Dunes: their geomorphology,
stratigraphy and soils as a basis for habitats of Banksia woodlands. Journal of the Royal
Society of Western Australia 71: 87–88
Specht RL, Yates DJ, Sommerville JEM and Moll EJ (1991) Foliage structure and shoot growth
in heathlands in the mediterranean-type climate of southern Australia and South Africa.
Ecologia Mediterranea 16: 195–207
Speck NH (1952) Plant Ecology of the Metropolitan Sector of the Swan Coastal Plain. MSc
Thesis, The University of Western Australia
Stewart GR, Pate JS and Unkovich MJ (1993) Characteristics of inorganic nitrogen assimila-
tion of plants in fire-prone Mediterranean-type vegetation. Plant, Cell and Environment 16:
351–363
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.