Our Valuable Native Grasslands, Better Pastures Naturally
Proceedings of the Second National Conference of the Native Grasses Association

Guiding the delimitation of local provenance for Australian native grasses.

C. M. Waters

Rangelands Research Officer, NSW Agriculture, Trangie 2823

Abstract

The provenance of a seed source not only describes its point of collection but can also reflect its evolutionary development.

Here, local provenance material may be the best site-adapted material and the use of non-local material, or less `fit' ecotypes may result in establishment failure, long-term mortality or disruption to the surrounding ecosystem. No guidelines for the definition of a provenance currently exist for any Australian native grass, providing a major obstacle for the broad-scale use of Australian native grasses.

Understanding the evolutionary forces such as natural selection, gene flow and genetic drift as well as the factors that operate to drive these forces, can provide clues to the definition of a provenance boundary. If native grasses are to retain their adaptive and low input advantages over the use of exotic species then understanding the issues of provenance provides the key to their successful and widespread use.

1. Introduction

Provenance is a term used to describe something's origin or source and in the case of seed as its "geographic place of origin or seed" (Loch and Whalley 1997). This term is also broadened to describe the patterns of variation exhibited by a species over its range reflecting its evolutionary history (Coates and van Leeuwen 1996). Variation within populations of Australian native grasses is used to continually re-appraise and truncate complexes. For example, the recent division of Danthonia linkii into Austrodanthonia bipartita and Austrodanthonia fulva (Linder 1997). However frustrating these name changes are to those of us working with these species the importance of the delimitation of taxon is unquestionable. The adaptive significance of this variation amongst populations remains largely unrecognised as this fine level of detail appears logistical impossible to document and is of perceived limited use. It is this perception that is questionable. The result has been an almost fashionable philosophical debate on when it is or is not appropriate to use local provenance material in revegetation programs. This debate is largely based on theoretical arguments and offers little in the way of practical advise on how to synthesise knowledge to make an informed decision about the genetic integrity of seed sources.

Arguments in favour of using local provenance material centre around the idea that seed of local origin is best adapted to the site. The debate extends to suggest that when material is not of "local provenance" it may be mal-adapted to the target site and/or could result in genetic pollution of local populations, in extreme circumstances resulting in a reduction in local biodiversity. Given the goals of most native revegetation programs are to conserve or restore native populations that are self-perpetuating, of low maintenance (Wilson, 1996) and integrated with surrounding ecosystems (Coates and van Leeuwen 1996), concerns that the use of inappropriate genetic material will result in further ecosystem damage are alarming. Apart from these negative consequences the major argument against using local provenance material in reseeding programs is that we are unable to easily define provenance boundaries or delimitate a provenance area. As a result there are no guidelines to assist in formulating seed collection strategies. This paper attempts to provide the guidance in developing these collection protocols through an understanding of the forces that have resulted in ecotypic development.

2. Why is provenance an important issue?

2.1 Biodiversity

Whilst we would all agree that biodiversity gives us the spice of life and there is broad public acceptance of the importance of the conservation of biodiversity at a species level, recognition that biodiversity is also essential to life itself is not so easily acknowledged (Vitousek 1997). If we consider the overall genetic architecture of a species in terms of within and between population variation, it is the between population differences that reflect the historical actions of natural selection in shaping local adaptation. The within-population variation that indicates the capacity of the population to evolve in response to future selection pressures. In other words intra-specific biodiversity provides insurance against environmental changes and a species future capability to adapt and survive.

Patterns of the genetic variation within a species are hierarchical in nature. The total genetic variation in a species can be organised as:

(i) Variation amongst geographic regions

(ii) Variation among stands or populations within regions

(iii) Variation among families (kinship groups) within stands

Variation among siblings within families (Millar and Libby 1989).

Because plants are sessile, and dispersal of seed or pollen over large distances occurs infrequently, local differentiation among plant populations is common. Within any species there may be distinct geographical or ecological segregates called subspecies or varieties. This level of variation provides a useful source of material for our plant breeding programs. The term "ecotype" is used to subdivide a species into populations based on a distinct set of characteristics that appear to indicate that it is adapted to local or regional conditions. If we are serious about conservation of biodiversity, we need to preserve each of these levels of biodiversity. If the goal of revegetation is to re-create or restore native communities, this implies some commitment to replication of the genetic structure of the community and thus an understanding of each level of variation may be required. For the more consumptive industries, most landholders can settle for something less than this. A primary producer may have a revegetation goal to design a biologically viable and sustainable planting and thus would require an understanding of variation at a level among populations only. This goal would provide a basis on which to balance the risks of not using local provenance material against sourcing local material (Millar and Libby 1989; Jones 1997).

2.2 The "home side advantage"?

The arguments in favour of using local provenance material primarily revolve around the idea that local seed is the best site adapted material (van Leeuwen 1994; Loch and Whalley 1997). If populations have become locally adapted to soil and various other environmental characteristics, then introducing non-local material may result in reduced survival, growth, or reproduction relative to locally adapted material. Using such mal-adapted material may result in both short and long-term undesirable consequences. Material that poorly matches the restoration site may result in a short-term establishment failure, increasing the cost and time frame for the revegetation program. There are examples from within the forestry industry that illustrate this, for Pinus ponderosa (Squillance and Silen 1962) and Pinus contorta (Illingworth 1975). The death or reduced vigour of poorly adapted material may be gradual, occurring over longer periods of time. In particular, this can occur with long-lived species such as Pinus taeda (Zobel and Talbert 1984).

A commonly used method to demonstrate differences in adaptation between populations is to undertake a reciprocal transplant experiment. Here, for a number of years, two or more populations are grown out in their respective local or "native" habitats as well as in their reciprocal habitats. Various "fitness" traits are then measured e.g. probability of seedling emergence, early vegetative size, probability of flowering and total number of inflorescences produced. There have been a number of examples where this has demonstrated local adaptation (Nagy and Rice 1997, Hewitt 1988 and Waser and Price 1985) and others where no adaptation or superiority has been demonstrated (Gordon and Rice, 1998). There appears to be no documented studies for Australian native grasses and we lack time and resources to undertake such studies. A further limitation in this method is that it may be unclear from these experiments whether any initial fitness advantage of the local form is persistent over time as these experiments usually only run for a number of years.

The term `reduced fitness' may not simply apply to introduced plants but also to the hybrids between it and the local population (Coates and Leeuween 1997). This term is referred to as `outbreeding depression'. This has been illustrated in a study of Amphicarpaea bracteata (a selfing annual) where the crossing of two divergent populations resulted in progeny with lower seed yield than either parent (Parker, 1992). In this way wild areas surrounding the restored site may be contaminated, and the genetic composition of the native populations irreversibly altered.

In an ongoing common garden study involving nine sites throughout Australia the performance of different ecotypes of native and exotic grasses is being examined. At one location at Trangie (central New South Wales) a local ecotype of Austrodanthonia ceaspitosa (Dc1) exhibited 30% survival compared to 2407, a Tasmanian ecotype, that resulted in less than 10% survival. At a second location outside Canberra, Dc1 and 2407 recorded 88% and 50% survival respectively, suggesting the Tasmanian ecotype is not well adapted beyond its southern range and the Trangie ecotype may extend its range some 300 km to the south (Norton et al. 2001). At the Canberra site, two local ecotypes of Microlaena stipoides (Ms1 and Ms2) were compared. Here, production and survival rates were similar for the two ecotypes, however the recruitment rate of Ms1 was almost twice that of Ms2, the latter ecotype being derived from a basalt soil compared to Ms1 that grew on a granite based soil and illustrating a different scales at which ecotypic development may take place.

2.3 Provenance and the developing Australian native seeds industry

With few commercially available varieties of Australian native grasses (Waters et al 2000) the market relies heavily upon opportunistic harvesting of seed from wild stands (Mortlock 1999; Waters 1997; Loch and Whalley 1997). This reliance has created problems in seed quantity and quality. As seed supply is determined by local seasonal conditions, sourcing non-local seed is sometimes the only option. Within the southern and eastern states of Australia there appears to be a policy unwritten (G. Allen pers. comm; N.B. Bonney pers. comm; B. Meyers pers. comm) or otherwise (Mortlock 1999), to access and use local provenance material where ever possible. In particular, Landcare and community revegetation projects in Australia are creating a demand for seed that is collected locally and thus the genetic quality of native seed is of increasing importance (Mortlock 1999). These concerns are matched elsewhere in the world where the native grass seed industries are well established such as western United States and Canada (Jones and Johnston 1998). Internationally, the genetic composition of plant material for revegetation programs is of great interest to a range of land managers, scientists and policy makers (Jones 1997; Loch and Whalley 1997; Knapp and Rice 1994; Meyer and Monsen 1993). In Australia, the practical application of using local provenance material may lead to a compromise in revegetation activities through the use of exotic grasses, inappropriate native species or prohibited entirely. Whilst various state native vegetation Acts encourage revegetation and enhancement of native grasslands and grassy woodlands, in reality broad-scale revegetation activities are confined to tree planting and understorey vegetation remains largely ignored due to these difficulties in seed supply and issues of genetic integrity of seed sources.

3 Evolutionary causes of ecotypic development.

Understanding the evolutionary pathways involved in the development of differences within and between populations is a first step towards being able to guide seed collection protocols. Understanding the genetic basis for intraspecific phenotypic variability depends on knowing something about the potential influence of the environment in shaping phenotypes.

What causes local adaptation ?

Natural selection: Genetic variation is a reflection of environmental heterogeneity that exists within a species range. Because plants can not move far, they are forced to cope with environmental conditions that occur at the site in which they germinate. Natural selection will act, through differential survival and reproduction and sometimes results in the formation of ecotypes. As selection acts directly on phenotypic variation, the relative importance of genetic differentiation will depend on the relationship between genotype and phenotype (heritability). Factors such as soil and altitude gradients, climatic variation, disease and competition can all cause major morphological and physiological differences among populations. When selection is strong and consistent in one direction, the result can be local adaptation (Figure 1). Whereas ecotypic variation between populations represents the products of selection, within population genetic variation represents the future potential for selection. The effectiveness of selection is going to depend on the amount of genetic variation that exists within a population initially.

Figure 1. Evolutionary pathways responsible for the divergence in a common ancestral population to separate species: adapted from Coates and van Leeuwen (1997)

3.2 What prevents local adaptation?

Genetic drift: There are situations where local adaptation may not occur, for example, when the occurrence of a particular genotype on a site might be attributed to chance alone, and not to natural selection and therefore does not represent any adaptive superiority. This will lead to a random change in population gene frequencies known as genetic drift. Genetic drift appears to only be of significance in small populations or isolated populations, where alleles may be easily extinguished.

Bogus' adaptations may also occur where a plant may have the appearance of local adaptive differences (it is locally wide spread), when none exist. For example, an immigrant into an open habitat with few competitors that can also self-fertilise will propagate and colonise the open habitat. Even though it rapidly spreads, this new sub-population of related genotypes is not necessarily better adapted to this open site than the surrounding population. This sub-populations dominance of the site reflects the action of chance colonisation, not the action of natural selection. Such founder individuals of colonising populations have only a small and probably non- representative sample of the parent population's gene pool.

Phenotypic plasticity: The capacity for a single genotype to express different phenotypes in different environments is called phenotypic plasticity. We often see grass responding to continually heavy grazing by developing a more prostrate growth habit. Phenotypic plasticity can be especially pronounced in plants (Rice and Knapp 2000).

Gene flow: Whilst natural selection and genetic drift can both lead to population divergence and spatial patterns in genetic variation, such differentiation is countered by another evolutionary force, that of gene flow. Gene flow is a function of the movement of pollen and seed between populations and acts to reduce the effects of genetic differentiation by effectively spreading genes through different populations (Baker 1951; Nagy and Rice 1997). Factors such as gene flow may effectively prevent the development of locally adapted populations. Self-pollinating species will tend to have little gene flow compared to out-crossing species (Whisenant 1999). Wind pollinated plants tend to have more rapid gene flows than do animal pollinated species. Under other conditions, gene flow may generate desirable new genetic combinations (Jones 1997). For example, in cross-pollinating species, short-term outbreeding depression could be offset by increased vigour through heterosis. In the longer-term continual segregation and natural selection could result in a better adapted ecotype (Rice and Knapp 2000;Loch and Whalley 1997).

Gene flow, founder effects and the ability of plants to acclimatise to a wide range of environments (phenotypic plasticity) may reduce the probability that local adaptation.

4 Delineating `seed collection zones'.

There are no clear recipes for describing seed collection zones and in many cases collection will require species-specific recommendations. However, as a first step towards describing seed collection zones, Coates and van Leeuwen 1996 suggest that consideration of five major factors can assist in minimising the risks of collecting material that is not site-adapted. They are, in order of complexity,

(i) Recognised morphological variation

(ii) Distribution patterns

(iii) Biogeographical and ecological information

(iv) Breeding system

(v) Genetic structure of the population

4.1 Morphological differences and the scale of ecotypic variation

The importance of recognising morphology differences between populations cannot be under estimated because it is these physical differences that can often reflect the effects of natural selection and thus adaptation. Understanding the scale of variation at a population level can give indicators to the existence of ecotypic development. For example, narrow scales of within population phenotypic variation may increase the likelihood of ecotypic variation where as broader scales may decrease its occurrence.

There are more than 750 species of Australian native grasses in about 180 different genera (Lodge and Groves 1990). Very few have been studied in any detail and as a result we have a limited understanding of the scales of variation between populations within species. Recent native grass breeding and selection programs have documented large phenotypic differences within species providing sufficient variation to allow selection of distinct and uniform cultivars of grasses (Mitchell et al.2001; Waters et al. 1997).

In a common garden study, fifteen native grass species collected from 51 sites throughout western New South Wales and south-west Queensland and grown together for 3 years. A principal component analysis revealed a strong relationship between site of origin and plant morphological and floristic characteristics providing strong evidence for distinct ecotypic separation for at least 10 of these species. Marked clustering was strongly influenced by variation in plant size among species. Groupings for four of these species are shown in Figure 2.

Figure 2. Plots of plants grown in a common garden on the first two principal components derived from analysis of morphological and floristic characteristics for Dgcc (Digitaria coencicola), Dchs (Dichanthium sericeum), Dacp (Danthonia caespitosa) and Etpa (Enteropogon acicularis). Plants collected from the same location are identified by the same letter. These scatter diagrams reveal marked clustering according to site.

Despite the observed high degree of variation in our native grasses, some species do appear to have a wide adaptation to a range of environments. For example, Astrebla pectinata cv. Turanti, originated from a Camerons Corner accession (far North Western New South Wales) but performed well some 650 km east at a number of sites with 200 mm higher annual rainfall than its origin (Waters et al. 1998). Similarly, another native grass, Microlaena stipoides cv. Griffin which originated in Canberra grew well 800 km north in Armidale (Whalley and Jones 1995). For other species the scales of variation can be much finer. For example, Magcale-Macandog (1994) found Microlaena stipoides growing in a patch of Lolium perenne (perennial rye grass) to be genetically distinct from M. stipoides growing in a patch of Poa pratensis (Kentucky bluegrass) in the same paddock. This micro-evolutionary differentiation occurring over a very short period of perhaps 30 years (Groves and Whalley 2001)

4.2 Species distribution patterns

Mortlock (2000) suggests three factors are important in assessing the influence of distribution patterns on local adaptation, the extent or scale of distribution, the population density and the fragmentation of the landscape. Populations that are continuous in range and share a common, non fragmented environment, the trend is to have less differentiation between populations because gene flow is generally high with no natural barriers. For example, Mitchell grass Astrebla lappacea is a widely distributed species that occupies a broad geographic range from northern NSW, through western Queensland and the Northern Territory. Characteristically this species is found on grey cracking clay soils. Collections from almost 278 different populations of A. lappacea were grown in a common garden at Walgett and morphological characteristics examined. Here, no relationship between site of origin, plant morphological and floristic characteristics was found. This suggests that A. lappacea exhibits adaptation to a wide range of environments and provenance boundaries might be large (Figure 3). On the other hand, disjunct populations will tend to exhibit strong ecotypic development because physical barriers, effective in preventing gene flow, favour the development of ecotypes (Knapp and Rice 2000).

Figure 3. Plots of plants grown in a common garden on the first two principal components derived from analysis of morphological and floristic characteristics for Curly Mitchell grass (Astrebla lappacea) and Barley Mitchell grass (Astrebla pectinata). Plants collected from the same location are identified by the same letter. Scatter diagram reveals no marked clustering according to site of origin suggesting a distinct lack of ecotypic development.

4.3 Reproductive systems

Because gene flow is a function of pollen dispersal, the breeding system of a plant will have a major influence on both the intensity and spatial scale of local adaptation (Rice and Knapp 2000). In cross-pollinating, or outcrossing species, gene flow is high because outcrossing promotes more extensive movement of pollen (and gene flow) and thus local adaptation is unlikely to occur (Jain and Bradshaw 1966; Slatkin 1981). If a species is self-pollinating, genetic differences among populations may exist because of a lack in gene flow.

We have only a very limited understanding of the breeding systems in native Australian grasses. Whilst the relative proportion of selfing to crossing populations is likely to differ between species, there appears that at least the proportion of selfing populations is high, perhaps as much as 80% (Groves and Whalley 2001). We also know that apomixis is common in Australian grasses. For apomictic plants pollination occurs however fertilisation does not. Here, viable seeds are produced asexually and the embryo will usually have a genotype identical to that of the parent. In this way apomixis excludes variability in offspring and offers an ideal mechanism for the rapid proliferation of `fit' genotypes in environments that are uniform Groves and Whalley). Apomixis will slow down gene flow between populations so a high level of local adaptation can develop. Sexual reproduction results in the production of polymorphic offspring capable of exploiting variable environments. If a species is able to evolve a balanced sexual and apomictic breeding system, the high cost of sexual reproduction can be minimised and some degree of adaptive polymorphism can be retained (Groves and Whalley 2001). This `flexibility' in breeding system offered by apomixis is common in the Poaceae family world-wide and is likely one of the keys to the ecological success of this family through different climates.

At first this would appear not to make ecological sense, since Australian landscapes are patchy in nature with a highly variable climate and it would be better perhaps to have a breeding system where variability was encouraged. Encouraging variability through sexual reproduction would therefore provide the capacity for a grass to respond to a fluctuating environment. However, most perennial native grasses have evolved mechanisms to cope with surviving in harsh landscapes, and can be classed as `drought-resisting' or `drought-evading' (Lazarides 1970). So Australian grasses can produce large numbers of a specific genotype that is successful in a certain environment, if the environment changes dramatically then a switch to sexual reproduction can be made. Sometimes things can go horribly wrong when the reproductive system is not well balanced and the development of apomixis can lead to an `evolutionary dead-end'. For example, Bothriochloa biloba and Dichanthium setosum, both vulnerable Australian native grasses, produce multiple embryos that lead to abortion and low seed set (Yu 1999; Yu et al. 1999).

4.4 Biogeographical and ecological factors

Generally, most people don't have a hard time recognising broad geographic gradients such as changes in altitude, rainfall and day-length but often the role they play spatially and temporally in species differentiation is not so readily recognised.

Genetic variation is a reflection of environmental heterogeneity that exists within a species range. Because plants can not move far, they are forced to cope with environmental conditions that occur at the site in which they germinate. Factors that help to create these adaptations include climatic conditions (e.g. rainfall probability, frost tolerance, germination temperatures), edaphic conditions (e.g. soil texture, water holding capacity, chemical composition) as well as ecological processes (mycrorrhizal associations, pollinator abundance, seed dispersers). For example, Hodgkinson and Quinn 1978 found that in cool moist environments such as those of Tasmania and Victoria Austrodanthonia ceaspitosa flowering responded to changes in day length. Conversely, in the hot, semi-arid environments found in western New South Wales the development and control of reproduction appears to be more plastic to allow for opportunistic growth and reproduction in response to soil moisture availability. It would be sensible to suggest therefore that moving the Tasmanian types to western NSW and vice versa would result in reduced performance.

4.5 Identifying a genetic basis for population differences

Today the diversity of molecular markers available for studying the genetic structure of different populations is considerable (Krauss et al 1999; Coates and van Leeuwen, 1996). It should be stressed that molecular markers are only able to resolve patterns of geographic genetic variation within a species and do not infer adaptive differences. As a result they provide a potentially useful tool for definition of provenance areas when used in combination with other information about the species such as distribution patterns, reproductive system and geographic information. There are numerous examples of this from rare and endangered flora in the south-west of Western Australia (Coates 1992; Coates and Sokolowski 1992).

For other more widely spread native species such as Acacia holosericea isozyme studies have revealed two distinct populations that match different biogeographic regions, northern Australian arid and tropical zones (van Leeuwen 1994). As these populations also had distinct morphological characteristics reappraisal has led to the creation of a new taxon A. colei that occurs in the arid areas (Maslin and Thomson 1992). For another Western Australian species broadly adapted to the south southern tip of this state Eucalyptus diversicolor, displayed little differentiation between main forest populations (Coates and Sokolowski 1989). This could be explained by other factors. For example, the distribution of this species, though limited in range, was continuous with no significant barriers to gene flow and therefore likely high levels of outcrossing between populations (Coates and van Leeuwen 1996) suggesting that provenance areas for this species are likely to be large. As such combining genetic information with other factors can help to resolve provenance areas.

4.6 Attempts at delineating provenance areas

Where a species is still in relative abundance seed can be collected in close proximity to the restoration site. Here, seed collection should ensure adequate sampling of the existing genetic variation. The more plants from which seed is collected, the greater the chance of obtaining good population representation. Collection guides that assist in retaining biodiversity at this scale have been documented (Knapp and Rice 1994; Mortlock 2000).

When a species is not present on the restoration site, collections need to be made elsewhere. Here, a simple collection radius can be drawn up around the restoration site to define a collection zone. These have been done elsewhere in the world for many tree and shrub species (Jones and Johnston 1989: Millar and Libby 1989). Linhart (1995) advocates collection from within a 100m radius from the restoration site for herbs and within 1 km of the site for woody plants. Geographic distance alone is an unreliable guide and can be supplemented by matching site conditions between the collection point and the restoration area (Jones and Johnston 1998). Matching can be done to reflect a particular combination of climatic, microclimatic, topographic and edaphic variables (Millar and Libby 1989, Meyer and Monsen 1993). Regions of homogeneous environments rarely follow contours on a map. Here, the establishment of zones based on broad ecological and geographical similarities can be useful. Between these zones the transfer of grass is avoided. The US Forest Services users "seed zones" to guide conifer seed collection for revegetation (Kitzmiller 1990). According to these guidelines replanting must be from stock native to the seed zone. Data on the spatial patterning of genetic variation is helpful for delineating "seed zones" but not yet available for Australian native grasses. Until this data is collected it may be reasonable to create preliminary seed zones based on broad biogeographical (or major botanical) divisions of a State.

Coates (2000) suggests that conservation units based on DNA testing and used by conservation biologists to define genetically distinct populations or clusters of populations may be useful in defining a collection area. Specifically `Evolutionary Significant Units (ESU) reflect a historically isolated, independent set of populations that warrant separate management for conservation (Ryder 1986).

Conclusions

The value of using native grasses in revegetation programs lies in both their low input advantages and in their ability to withstand our variable Australian climate (Loch and Whalley 1997; Waters et al. 1997; Wilson 1996). Their role in the maintenance of biodiversity and ecological integrity of agricultural areas has also been widely promoted (Dowling and Garden 1991; Foreman 1995; Mortlock 1999). If native grasses are to retain these advantages over the use of exotic species then the issue of the genetic integrity of seed sources remains at the foundation of this argument and therefore must be investigated. Given we have little information on the spatial patterning of genetic variation in Australian native grasses, at least, a cautionary approach should be adopted (Loch and Whalley 1997). The lack of clear guidelines for the delimitation of local provenance in Australian native grasses need not preclude the consideration of local adaptation amongst populations as the factors responsible for their development are understood.

References

Baker, H.G. (1951). Hybridisation and natural gene flow between higher plants. Biological Review. 26 pp 302-337.
Coates, D.J. (1992). Genetic consequences of a bottleneck and spatial genetic structure in the triggerplant Stylidium coroniforme (Stylidiaceae). Heredity 69 pp. 512-520.
Coates, D.J. (2000). Defining conservation units in a rich and fragmented flora: implications for the management of genetic resources and evolutionary processes in south-west Australian plants. Australian Journal of Botany. 48 pp329-339.
Coates D.J and van Leeuwen S.J. (1996). Delineating seed provenance areas for revegetation from patterns of genetic variation. In: Proceedings of the second Australian workshop on Native Seed Biology for Revegetation. Eds. S.M.Bellairs and J.M. Osborne, 11-12 October, 1996, Newcastle. Australian Centre for Minesite Rehabilitation Research. pp 3-14.
Coates, D.J. and Sokolowski, R.E. (1989). Geographic patterns of genetic diversity in Karri (Eucalyptus diversicolor F. Muell.). Australian Journal of Botany. 37 pp. 145-156.
Coates, D.J. and Sokolowski, R.E. (1992). The mating system and patterns of genetic variation in Banksia cuneata A.S. George (Proteaceae) Herdity 69, pp11-20.
Dowling, P.M and Garden, D.L. Eds. (1991).Proceedings of the Native Grass Workshop, Dubbo. 1990.
Foreman, P. (1995). Understanding the ecological value of native grasslands in the winter rainfall region of the Murray Darling Basin using natural history as a framework. In: Proceedings of the Community Grasses Project Workshop. Report No 39. (Johnstone Centre of Parkes, Recreation and Heritage, Albury pp. 16-24.
Gordon, D.R. and Rice, K.J. (1998). Patterns of Differentiation in Wiregrass (Aristida beyrichiana): Implications for restoration efforts. Restoration Ecology. 6 (2). pp. 166-174.
Groves, R.H and Whalley, R.D.B. (2001). Ecology of the major Australian grass genera and the grasslands in which they occur. In: Flora of Australia. Vol 43 (in press) .
Hewitt, G.M. (1988). Hybrid zones - natural laboratories for evolutionary studies. trends in Ecology. 3: pp158-167.
Hodgkinson, K.C. and Quinn, J.A. (1978). Environmental and genetic control of reproduction in control of reproduction in Danthonia ceaspitosa populations. Australian Journal of Botany. 26 pp. 351-364.
Illingworth, K. (1975). Lodgepole pine provenance research and breeding in British Colombia. In: Management of lodgepole pine ecosystems. Cooperative Extension Service Washington State University. Vol. 1: p 47-67.
Jain, S.K. and Bradshaw, A.D. (1966). Evolutionary divergence among adjacent plant populations. I. The evidence and its theoretical analysis. Heredity. 21. Pp407-441.
Jones, T.A. (1997). Genetic considerations for native plant material. In: Proceedings: Using seeds of Native Species on Rangelands. Society for Range Management 50^th Annual Meeting, Rapid City, South Dakota. February 16-21. pp. 22-25.
Jones, T.A and Johnston, D.A. (1998). Integrating genetic concepts into planning rangeland seedings. Journal of Range Management. 51: pp 594-606.
Kitzmiller, J.H. (1990). Managing genetic diversity in a tree improvement program. Forest Ecology and Management. 35 pp.131-149.
Knapp. E.E. and Rice, K.J. (1994). Starting from Seed. Genetic Issues in Using Native Grasses for Restoration. Restoration and Management Notes. 12:1. pp 40-45.
Krauss, S.L., Hood, P., Zawko, G. and Mittner, J. (1999). Recent advances and new genetic tools for the delineation of provenances. In: Proceedings of the Third Australian Workshop on Native Seed Biology for Revegetation. Ed. C.J. Asher and L.C. Bell. Australian Centre for Minesite Environment Research. pp. 13-23.
Linder, H. P. (1997). Nomenclature corrections in Rytidosperma complex (Danthonieae, Poaceae). Telopea 7(3): pp 269-274.
Linhart, Y.B. (1995). Restoration, revegetation, and the importance of genetic and evolutionary perspectives. In: B.A. Roundy, E.D. McArthur, J.S. Haley and D.K. Mann. Proceedings of wildland shrub and arid land restoration symposium. USDA Forest Service. General Technical Report. INT-GTR-315. Pp. 271-287.
Loch D. and Whalley R.D.B. (1997). Native Grassland Rehabilitation: The Place of Provenance Seed. In: Proceedings of the First workshop of the Australian Native Grass and Legume Seed Industry Association Inc., 22 Oct 1997. Brisbane. pp 59-69.
Lodge, G.M. and Groves, R.H. (1990). The domestication and agronomy of native grasses. In: Proceedings of the Native Grasses Workshop.Ed. P.M. Dowling and D.L. Garden. Dubbo, 16-17 October, 1990. Pp 73-81.
Maslin, B.R and Thompson, L.A.J. (1992). Reappraisal of the taxonomy of Acacia holosericea, including the description of a new species, A. colei, and the reinstatement of A. neurocarpa. Australian Systematic Botany. 5 pp. 72-743.
Millar, C.I., and Libby, W.J. (1989). Disneyland or Native Ecosystem: Genetics and the Restorationist. Restoration and Management Notes 7:1. pp18-24.
Mitchell, M.l., Koen, T.B. Johnston, W.H. and Waterhouse, D.B. (2001). LIGULE: An evaluation of indigenous perennial grasses for dryland salinity management in south-eastern Australia. 2. Field performance and the selection of promising ecotypes. Australian Journal of Agricultural Research, 52, pp. 351-365.
Mortlock, W. (1999). Native seed in Australia. A survey of collection, storage and distribution of native seed for revegetation and conservation purposes. A report to the FloraBank project. Canberra.
Mortlock, W. (2000). Guidelines 10, Seed collection ranges for revegetation. Flora Bank.
Meyer, S.E. and Monsen, S.B (1993). Genetic considerations in propagating native shrubs, forbs and grasses from seed. In: Proceedings of the Western Forest Nursery Association. 1992. USDA Forest Service General Technical Report. RM -221. pp. 47-54.
Nagy, E.S. and Rice, K.J. (1997). Local adaptation in two subspecies of an annual plant: Implications for migration and gene flow. Evolution. 51 (4). pp 1079-1089.
Norton, M.R., Garden, D., Waters, C., Whalley, R.B.D., Friend, D., Mitchell, M., Kobelt, E., Aurich, G and Stanford, P. (2001). The native and Low input Grasses Network - Multi-site testing. Final report prepared for MLA. ISBN 0 7347 1296 0.
Parker, M.A. (1992). Outbreeding depression in a selfing annual. Evolution, 46(3). pp. 837-841.
Rice, K.J. and Knapp, E.E. (2000). Evolutionary factors affecting the probability of local adaptation or should we expect to see ecotypes behind every rock ? In 2^nd. Interface Between Ecology and Land Development in California. Ed. J.E. Keeley, M. Baer-Keeley and C.J. Fotheringham. U.S. Geological Survey Open-File Report 00-62. Pp. 221-226.
Ryder, O.A. (1986). Species conservation and systematics: the dilemma of subspecies. Trends in Ecology and Evolution. 1 pp9-10.
Slatkin, M. (1981). Estimating gene flow in natural populations. Genetics. 99. Pp. 323-335.
Squillance, A.E and Silen, R.R. (1962). Racial variation in ponderosa pine. Forest Science Monograph. 2. p 27.
van Leeuwen, S.J. (1994). Justification for the use of provenance seed in the rehabilitation of disturbed sites. In: Proceedings Revegetation of Mine sites Using Appropriate Species. Workshop at Third International Conference on Environmental Issues and Waste Management in Energy and Mineral Production. Ed. K.A. O-Connel and D.R. Barrett. Mulga Research Centre, Bently pp. 27-36.
Vitousek P.M, D'Antonia C.M., Loope, L.L., Rejmanek, M., and Westbrook, R. (1997). Introduced species: a significant component of human caused global change. New Zealand Journal of Ecology. 2 (1); pp.1-16.
Waser, N.M and Price, M.V. (1985). Reciprocal transplants with Delphinium nelsonii (Ranunculaceae): evidence for local adaptation. American Journal of Botany. 72. pp 1726-1732.
Waters, C.M., Loch, D.S. and Johnston, P.W. (1997). The role of native grasses and legumes for land revegetation in central and eastern Australia with particular reference to low rainfall areas. Tropical Grasslands 31, pp. 304-310.
Waters, C., Whalley, W. and Huxtable, C. (2000). Grassed-up: Guidelines for revegetating with Australian native grasses.
Whalley, R.D.B. and Jones, C.E. (1995). Microlaena (Microlaena stipoides). Plant Varieties Journal 8 (1), pp.27-28.
Whisenant, S.G. (1999). Repairing Damaged Wildlands- A process-orientated, Landscape-Scale approach. 1. Biological Conservation, Restoration, and Sustainability, University of Cambridge, United Kingdom.
Wilson, A.D. (1996). Native and low-input grasses for pastoral and marginal cropping land. New Zealand Journal of Agricultural Research 39, pp 465-469.
Zobel, B. and Talbert, J. (1984). Applied Forest Tree Improvement. J. Wiley and Sons. New York. p 505.