76 Protection Quarterly Vol.13(2) 1998 benefit would be captured by wool and Campbell, M.H. (1977). Assessing the area Outline. Unpublished mimeograph, lamb producers who operate in that part and distribution of serrated tussock CSIRO Division of Plant Industry, Can- of the Australian wool and lamb indus- ( trichotoma), St. John’s wort berra. tries represented by the study area. (Hypericum perforatum var. angusti- Prograze (1995). NSW Agriculture. It must be emphasized that these re- folium) and sifton bush (Cassinia Vere, D.T., Sinden, J.A. and Campbell, sults are preliminary because both the arcuata) in New South Wales. Depart- M.H. (1980). Social benefits of serrated RPM and econometric modelling compo- ment of Agriculture New South Wales tussock control in New South Wales. nents require further refinement. In rela- Technical Bulletin No. 18. Review of Marketing and Agricultural Eco- tion to the RPM, further work is required Campbell, M.H. (1987). Area and distribu- nomics 48, 123-38. on refining the soils and rainfall digitized tion of serrated tussock (Nassella Vere, D.T., Auld, B.A., Auld, J.A. and data and to incorporate elevations into the trichotoma (Nees) Arech.) in New South Campbell, M.H. (1993). ‘Economic as- GIS model so as to be able to determine Wales, 1975 to 1985. Plant Protection sessments of serrated tussock (Nassella arable and non-arable country. Also, the Quarterly 2, 161-4. trichotoma) as a pasture weed. Weed livestock feed energy requirements in the Cousens, R. (1985). A simple model relat- Technology 7, 776-82. model are preliminary values as the study ing yield loss to weed density. Annals of Vere, D.T., Griffith, G.R. and Bootle, B.W. is awaiting the inclusion in the model of Applied Biology 107. (1994). The development and applica- energy requirements calculated by the Crean, J. (1996). New South Wales wool tion of a quarterly econometric model GRAZFEED model. Further work is nec- and sheepmeat budgets. In ‘Farm of the Australian prime lamb industry. essary in refining the econometric model’s Budget Handbook’. (NSW Agricul- NSW Agriculture, Economics Services wool industry specification and in the in- ture). Unit, Agricultural Economics Bulletin tegration of this component into the other Denne, T. (1988). Economics of nassella No. 11. livestock industry models. tussock (Nassella trichotoma) control in Vere, D.T., Jones, R.E. and Griffith, G.R. Despite the recognized deficiencies in New Zealand. Agriculture, Ecosystems (1997a). An integrated economic meth- the analysis, the study represents a signifi- and Environment 20, 259-78. odology for evaluating the impacts of cant economic contribution to the study of Edwards, G.W. and Freebairn, J.W. (1982). weeds in agricultural production sys- serrated tussock. The 1997 survey repre- The social benefits from an increase in tems and the farm and industry ben- sents the most accurate collection of data productivity in a part of an industry. efits of improved weed control. Techni- on the extent and distribution of serrated Review of Marketing and Agricultural Eco- cal Series No. 2, CRC for Weed Man- tussock in New South Wales. Having nomics 50, 193-210. agement Systems. these data incorporated into a GIS model Labys, W.C. (1984). ‘Commodity Models Vere, D.T., Jones, R.E. and Griffith, G.R. has made this information far more ame- for Forecasting and Policy Analysis’. (1997b). Evaluating the farm and indus- nable to modelling, particularly for incor- (Croom Helm, London). try impacts of weeds and the benefits of poration into an economic analysis. McDonald, W. (1996). Matching pasture improved weed control in agricultural production to livestock enterprises in production systems. Plant Protection References the Northern Tablelands and North Quarterly 12, 145-50. Alston, J.M. (1991). Research benefits in a West Slopes and Upper Hunter. Agnote multimarket setting: a review. Review of DPI/130, NSW Agriculture. Marketing and Agricultural Economics 59, Moore, A.D., Donnelly, J.R. and Freer, M. 23-52. (1990). The GRAZPLAN Model: An

that invade native vegetation usu- The biology of Nassella and Achnatherum species ally adversely affecting the survival of the indigenous flora. A potential impact of en- naturalized in Australia and the implications for vironmental weeds is a loss of biodiversity and a decrease in aesthetic value. management on conservation lands Stipoid grasses (in particular, N. neesiana and N. trichotoma) have proven to M.R. GardenerA and B.M. SindelB be difficult to control and have continued Division of BotanyA and Division of Agronomy and Soil ScienceB, School of to spread in conservation lands. They are Rural Science and Natural Resources and CRC for Weed Management successful because they have many bio- Systems, University of New England, Armidale, NSW 2351, Australia. logical traits which allow them to out- compete native vegetation. Stipoid grasses Summary Introduction generally invade plant communities Several species of Nassella and Several species of Nassella (Barkworth which are already highly degraded and Achnatherum are weeds of both conser- 1990) and Achnatherum (Barkworth 1993) have a history of disturbance (G. Carr per- vation and pasture lands. These species have become naturalized in Australia. sonal communication), and lands with have proven to be difficult to control and They include Nassella charruana (Arech.) higher fertility soil often previously used have continued to spread since their in- Barkworth, N. hyalina (Nees) Barkworth, for grazing or farming. These communi- troduction to Australia. The impact of N. leucotricha (Trin. & Rupr.) Pohl, N. ties may originally have been grasslands these species on conservation lands in- megapotamia (Spreng. ex Trin.) Barkworth, or grassy woodlands. Some conservation cludes a perceived drop in biodiversity N. neesiana (Trin. & Rupr.) Barkworth, N. areas in Victoria with significant inva- and a decrease in aesthetic value. Rea- trichotoma (Nees) Hackel ex Arech., sions of stipoid grasses include the Derri- sons for the ability of these species to Achnatherum caudatum (Trin.) S.W.L. mut Grasslands Reserve, Organ Pipes Na- out-compete native vegetation include Jacobs & J.Everett and Achnatherum tional Park and Southern Plenty Gorge. effective dispersal mechanisms, the pro- brachychaetum (Godr.) Barkworth. These Anecdotal evidence suggests that there duction of large amounts of aerial and species are referred to as stipoid grasses. is a drop in biodiversity in stipoid grass- clandestine seeds and large long-lived They are considered both environmental dominated grasslands because litter from seedbanks. Management strategies must and pasture weeds. Carr et al. (1992) de- the tall tussocks accumulates in the inter- take these factors into account. fines environmental weeds as exotic tussock spaces and excludes shade Plant Protection Quarterly Vol.13(2) 1998 77 intolerant species. However, thick stands to be achieved. Do we try to eradicate flowers (self-fertilized) which are ex- of undisturbed Themeda triandra Forssk. stipoid grasses or accept them as a perma- tremely variable and morphologically dif- have a similar inhibitory effect on other nent part of that community and try to ferent from those borne in the panicles. species (Stuwe and Parsons 1977). The di- manipulate the system to favour the na- Seeds from these flowers, known as versity of cryptogams (e.g. mosses, li- tive components? The aim of this paper is cleistogenes, originate from nodes on the chens, bryophytes) is thought to decline in to discuss some of the biological attributes flowering stem and are concealed under stipoid grass-dominated grasslands be- of Nassella and Achnatherum species which the leaf sheaths. In N. neesiana there is a cause the mosaic of substrates such as have resulted in their proliferation on con- progressive reduction in inflorescence rocks and bare soil becomes covered with servation lands. length and number of spikelets and floral litter (V. Stajsic personal communication). parts from panicle spikelets to spikelets at Quantitative studies are needed to com- Dispersal the base of the tillers (Connor et al. 1993). pare diversity of stipoid grass-dominated Long distance dispersal of the stipoid Five of the eight stipoid grasses in this grasslands and adjacent native remnants. grasses is by adhering to the coats of ani- study have axillary cleistogenes and four If there is reduced diversity, does it result mals, clothing or machinery via a sharp produce basal cleistogenes on the lowest from general degradation or is it specific callus at the end of the seed (Table 1). node of the flowering tiller (Table 1). to stipoid grass-dominated grasslands? Sheep can carry seed of N. neesiana in their Nassella neesiana and N. trichotoma have Likewise, what effect do stipoid grasses wool for at least 166 days (Gardener un- the potential to produce huge numbers of have on the diversity of vertebrate and in- published data). Thus there would be am- viable panicle seeds. In a dense, ungrazed vertebrate species? The striped legless liz- ple opportunity for dispersal over large infestation, seed production of N. neesiana ard (Delma impar) appears to prefer dense distances. Seeds probably also adhere to ranged from 1584 to 22 203 seeds m-2 intact swards of native tussock grasses other animals such as kangaroos and rab- (Gardener et al. 1996). These differences such as T. triandra and Austrostipa ssp., but bits. were correlated with the amount of spring can also use exotic species such as Phalaris Nassella trichotoma has tumbling inflo- rainfall, the former being in a drought spp. (Osbourne et al. 1993). How well do rescences which detach at the base, and year and the latter being in an above aver- they exist in degraded stipoid grass- can be windblown for up to 10 km dis- age year. Fluctuation in seed production dominated grasslands? persing seeds as they go (Campbell 1982). can be attributed mainly to the change in Management of weeds in natural sys- Jarava plumosa (Spreng.) S.W.L.Jacobs and number of flowering tillers produced per tems is often more difficult than in pas- J.Everett can also be blown by wind but unit area. Similar variability exists in seed tures or crops because there can be many unlike most stipoid grasses the back of its production of N. trichotoma (Healy 1945). more desirable species that need to be pre- lemma is also covered with long hairs In a heavily infested paddock in New Zea- served. Different species have varying re- (Burkart 1969). Heavy seeds without such land he found an average of 340 000 viable quirements for persistence such that one hairs to assist in wind dispersal, e.g. N. seeds m-2, whereas Campbell (1977) esti- management strategy is unlikely to cover neesiana, were found a maximum of 2.8 m mated N. trichotoma produced 93 000 all species. The aim of a study by Phillips away from the parent plant (Gardener un- seeds m-2 in Australia. and Hocking (1996) was to distil a set of published data). Even if panicle seed production could treatments that would allow a total re- While the primary dispersal mecha- somehow be reduced or stopped alto- placement of N. trichotoma by T. triandra on nism for N. trichotoma is wind, animals can gether, damaged tillers may still produce the conservation-listed western basalt inadvertently act as a secondary disperser cleistogenes. Damaged flowering tillers of plains of Victoria. Is this achievable in the after ingesting seeds. Campbell (1977) re- N. neesiana produced the same number of long term? Using the method detailed in ported the excretion of 4600 seeds per cleistogenes as undamaged tillers (Gar- Phillips and Hocking (1996), N. trichotoma- wether in the 10 days after grazing in- dener unpublished data). It appears that if dominated grasslands were replaced by fested paddocks. Healy (1945) found 70% N. neesiana plants are sprayed with a mix- dense swards of T. triandra after four years seed germination from cows which had ture of paraquat (Gramoxone) and flu- (C. Hocking personal communication). Is been fed N. trichotoma. On the other hand, propanate (Frenock) just after tiller elon- this T. triandra dominance self perpetuat- very few seeds of N. neesiana survived pas- gation, the basal cleistogenes still mature ing? What will prevent future recruitment sage through the guts of cows (e.g. 0.5% of whereas most of the panicle seeds and ax- from the presumably long-lived seed bank panicle seeds and 2.7% of cleistogenes re- illary cleistogenes do not. Dyksterhuis of N. trichotoma in these restored areas? covered from the faeces of Angus steers (1945) found that clipping N. leucotricha Undoubtedly, a dense canopy of T. were viable). All seeds had passed twice weekly to 4 cm above ground level triandra inhibits the germination and sur- through the animals within 4 days (Gar- reduced the production of basal cleisto- vival of N. trichotoma seedlings. However, dener unpublished data). genes but did not prevent it. other species may also be affected by this Stipoid grasses typically possess a Cleistogenes account for a considerable inhibition such as the endangered daisy, geniculate, hygroscopically-active awn. amount of total seed production in N. Rutidosis leptorrhynchoides F.Muell., which Straightening and twisting of the awn oc- neesiana (e.g. 21.5% (6079) of 28 282 seeds requires gaps larger than 50 cm in diam- curs with wetting and drying events, ena- m-2) (Gardener unpublished data). On eter for recruitment and survival (Morgan bling these seeds to move short distances average, N. neesiana produced 7.2 cleisto- 1997). Disturbance, such as the creation of and bury themselves in favourable micro- genes per flowering tiller (Gardener et al. gaps, is needed to maintain the frequency sites, thus increasing their probability of 1996). Sant et al. (1992) found that flower- of native species but it also promotes germination. ing tillers of A. brachychaetum produced up growth of the exotic species that make up to 18 cleistogenes whereas N. leucotricha is the bulk of the seedbank in degraded na- Reproduction capable of producing up to 12 (Dykster- tive grasslands (Lunt 1990a). Hence, in All stipoid grasses naturalized in Aus- huis 1945). managing the whole community, not just tralia are cool season perennials which In Argentina and the United States, A. stipoid grasses or the common native spe- grow during autumn and winter and brachychaetum is a serious weed in lucerne cies, we need to consider strategies that flower and set seed during the spring and pastures where cleistogenes play a major will maximize biodiversity. early summer. They reproduce through role in the plant’s persistence (Ares et al. Understanding the ecology of both the self-pollination or outcrossing and bear 1970). Dyksterhuis (1945) found that native and exotic flora is necessary if suc- seeds in terminal panicles (Connor 1987). cleistogenes in N. leucotricha are important cessful long term management goals are Some also have axillary cleistogamous in maintaining the species under adverse 78 Plant Protection Quarterly Vol.13(2) 1998 Table 1. A summary of biological attributes of naturalized stipoid grasses in Australia from the literature and current research. Question marks indicate unknown information. Dispersal Basal Stem Seed production Seedbank Seedbank Seed cleistogenes cleistogenes (seeds m-2) (seeds m-2) longevity (years) dormancy Nassella charruana ANoNo???? Nassella hyalina A No Yes ? ? ? ? Nassella leucotricha A Yes Yes ? 0–75 ? Yes Nassella megapotamia ANoNo???? Nassella neesiana A Yes Yes 1584–22 203 681–11 307 >6 Yes Nassella trichotoma W,A No No 93 000–340 000 1755–42 930 >13 Yes Achnatherum caudatum A Yes Yes ? ? ? Yes Achnatherum brachychaetum A Yes Yes ? ? ? Yes A = long distance dispersal via attachment to animals etc. W = wind dispersal.

Literature sources: Healy 1945, Ares et al. 1970, Campbell 1977, Campbell 1982, Joubert and Small 1982, Rosengurtt 1984, Kinucan and Smeins 1992, Vanauken 1997 and Gardener unpublished data. conditions, particularly under heavy After three years without seed input, and Small 1982, Vanauken 1997). Thus, grazing or burning, because they still de- the viable seedbank of N. neesiana (cleisto- seeds of N. leucotricha could remain veloped after flowering tillers were dam- genes and panicle seeds) in bare soil de- dormant and viable in a seedbank for a aged. Basal cleistogenes, which are often clined from 4676 to 1286 seeds m-2 (Gar- considerable time, which is uncharacteris- below the soil surface, have at least two dener unpublished data). If the seedbank tic of most grasses (Vanauken 1997). advantages. They are borne in a suitable continued to decline at this rate, it would habitat (as defined by the occurrence of take an estimated 11.4 years to reach 10 Implications their parent) and they are protected from seeds m-2. Bourdôt and Hurrell (1992) Preventing the dispersal of stipoid grasses extremes of climate, predation, grazing found that 17% of seeds buried at 5 cm in to new areas or ‘restored weed-free areas’ and fire. the ground were viable after six years. is very difficult because they effectively In New Zealand, N. neesiana was found There is also anecdotal evidence of germi- disperse by attaching to passing animals to develop basal cleistogenes in vegetative nation occurring after six years from con- (researchers included) and machinery. In tillers prior to flowering (Connor et al. tinually bared ground. the case of N. trichotoma, it is almost im- 1993). Similarly, Dyksterhuis (1945) found In New Zealand, seedbanks of N. possible to prevent wind dispersal from that plants of N. leucotricha commonly pro- trichotoma have been found to range from nearby infestations. duced cleistogenes before panicles ap- 1755 seeds m-2 in light infestations to The most important biological attribute peared on the plant. 42 930 m-2 in heavy infestations (Table 1) to consider when making management (Healy 1945). In Australia, 4784 seedlings decisions about weedy stipoid grasses Seedbanks m-2 have been recorded (Campbell 1958). is the potentially large, long-lived seed- One reason for the proliferation of stipoid Since only a small proportion of a seed- bank (at least for N. neesiana and N. grasses is their long-lived seedbanks. bank usually germinates at any one time, trichotoma). If herbicides or other forms of Studies on seedbank size and longevity the actual seedbank was probably much disturbance are used to kill the adult are only known for N. neesiana and N. larger. Seed of N. trichotoma can remain plants, it is likely that recruitment will oc- trichotoma. Using direct counts of viable viable for up to 13 years in the soil cur from the seedbank. Replacing the seeds (as opposed to germination counts (Campbell 1982). adult plant with competitive species such (Gross 1990)) seedbanks ranged from 681 In their countries of origin, the seed- as T. triandra will reduce the probability of seeds m-2 in a medium infestation to banks of stipoid grasses may not be so recruitment but it is unlikely that the 11 307 seeds m-2 in a dense infestation of large. A survey of seven populations of N. seedbanks will be sufficiently exhausted N. neesiana on the Northern Tablelands of neesiana on the Pampas Plains, Argentina, to prevent some regeneration. Seeds have NSW (Table 1) (Gardener unpublished revealed that there was almost no dormancy mechanisms that prevent them data). However, these values underesti- seedbank present (Gardener et al. 1997). all from germinating even if ideal condi- mated the total seedbank because only the Kinucan and Smeins (1992) found N. tions exist and make them persist for seeds loose in the soil (mainly panicle leucotricha to be common in Texan range- many years. seeds, but also some stem and basal lands (its native habitat), but it had a small In order to reduce the size of the seed- cleistogenes) were counted. It was found seedbank of 75 seeds m-2 or less. bank, seed production must be prevented. that in the discarded tiller bases (the soil Seeds often have dormancy mecha- However, unless plants are completely cores were passed through a 4 mm sieve nisms that allow them to persist for a long killed, it is difficult to prevent seed pro- to remove rocks and vegetation) there was time. Five of the eight stipoid grasses (Ta- duction because cleistogenes may still be an additional store of basal cleistogenes ble 1) are known to possess dormancy produced in some species. Seed produc- (e.g. a seedbank with 8335 panicle seeds mechanisms and it is likely that the others tion is variable depending on environ- m-2 had an additional 2963 basal cleisto- also do, since it is very common in the mental conditions but in N. neesiana, the genes m-2 or 35.5% extra). Stipeae. The hull or lemma around the seedbank can be maintained with low Ares et al. (1970) suggested that there seed may provide a barrier to gas and wa- seed production. The overwhelming evi- was a large store of A. brachychaetum seeds ter exchange and also mechanically re- dence that this species will persist no mat- (predominantly cleistogenes) in the soil strain the embryo. Before germination can ter what, has led us to recommend uti- with 80–100% of seedlings being of cleisto- occur, this barrier must be broken down. lizing it as a pasture plant on grazing gene origin. Similarly, about 50% of the Freshly harvested seed of N. trichotoma lands but this option is not available on seedlings of N. neesiana consistently arose and N. leucotricha both have after-ripening most conservation lands. from cleistogenes (Bourdôt and Hurrell requirements to prevent them germinat- 1992). ing in the summer after seed fall (Joubert Plant Protection Quarterly Vol.13(2) 1998 79 Conclusions Connor, H.E., Edgar, E. and Bourdôt, gamas del Uruguay. Boletin de la Evidence here suggests that Nassella and G.W. (1993). Ecology and distribution Universidad de la Republica 134, 1-28. Achnatherum species have biological char- of naturalized species of in New Sant, D.R.F., Eilberg, B.A. and Cruzate, acteristics that enable them to proliferate Zealand. New Zealand Journal of Agricul- G.A. (1992). Los granos cleistogenos in conservation lands. When management tural Research 36, 301-7. axilares del pasto puna (Stipa strategies are developed, they need to take Dyksterhuis, E.J. (1945). Axillary brachychaeta Gordon, Gramineae): su into account the biology of both the weeds cleistogenes in Stipa leucotricha and distribucion en las matas, en verano. and the species to be conserved. Lunt their role in nature. Ecology 26, 195-9. Revista de Investigaciones Agropecuarias, (1990) says that ‘ecologists face the formi- Gardener, M.R., Whalley, R.D.B. and INTA, Buenos Aires, Argentina 23, 77-84. dable task of devising disturbance re- Sindel, B.M. (1996). The failure of man- Stuwe, J. and Parsons, R.F. (1977). Themeda gimes that promote natives at the expense agement technology for reproductively australis grasslands on the basalt plains, of exotics’. Research is required to identify efficient grassy weeds: The Chilean Victoria: floristics and management ef- such management strategies. needle grass example. Proceedings of fects. Australian Journal of Ecology 2, the Eleventh Australian Weeds Confer- 467-76. Acknowledgments ence, pp. 243-6. (Victorian Weed Sci- Vanauken, O.W. (1997). Germination re- We would like to thank the Meat Research ence Society, Melbourne). quirements of aerial chasmogamous Corporation for funding the study on Gardener, M.R., Whalley, R.D.B. and florets and seeds of Nassella leucotricha which this paper is based. Sindel, B.M. (1997). The usefulness and (). Southwestern Naturalist 42, ecology of Chilean needle grass 194-200. References (Nassella neesiana) as a pasture in Ar- Ares, J., Soriano, A. and Eilberg, B.A. gentina. Unpublished report to the (1970). Mecanismos de invasion del CRC for Weed Management Systems, pasto puna (Stipa brachychaeta Godr.). 1. the Meat Research Corporation and the Caracteristicas de los diseminulos de la Weed Society of NSW. maleza. Revista de Investigaciones Gross, K.L. (1990). A comparison of meth- Agropecuarias, INTA, Buenos Aires, Ar- ods for estimating seed numbers in the gentina 7, 277-88. soil. Journal of Ecology 78, 1079-93. Barkworth, M.E. (1990). Nassella Healy, A.J. (1945). Nassella tussock. Field (Gramineae, Stipeae): revised interpre- studies and their agricultural signifi- tation and nomenclatural changes. cance. Bulletin of the New Zealand De- Taxon 39, 597-614. partment of Scientific and Industrial Re- Barkworth, M.E. (1993). North American search, No. 91, 90 pp. Stipeae (Gramineae): taxonomic Joubert, D.C. and Small, J.G.C. (1982). changes and other comments. Phyto- Seed germination and dormancy of logia 74, 1-25. Stipa trichotoma (Nassella tussock). Part Bourdôt, G.W. and Hurrell, G.A. (1992). 1. Effect of dehulling, constant tem- Aspects of the ecology of Stipa neesiana peratures, light, oxygen, activated char- Trin. & Rupr. seeds. New Zealand Jour- coal and storage. South African Journal nal of Agricultural Research 35, 101-8. of Botany 1, 142-6. Burkart, A. (1969). Flora ilustrada de En- Kinucan, R.J. and Smeins, F.E. (1992). Soil tre Rios (Argentina). (I.S.A.G., Buenos seed bank of a semi arid grass- Aires). land under three long-term grazing re- Campbell, M.H. (1958). The problem of gimes. American Midland Naturalist 128, serrated tussock in the Central Table- 11-21. lands of NSW. Australian Agrostology Lunt, I.D. (1990). The soil seed bank of the Conference, Armidale, CSIRO, Mel- long-grazed Themeda triandra grassland bourne, Volume 1, Part 2, Paper 73. in Victoria. Proceedings of the Royal Soci- Campbell, M.H. (1977). Serrated tussock. ety of Victoria 102, 53-7. Part 1: Life history, identification. NSW Morgan, J.W. (1997). The effect of grass- Department of Agriculture Division of land gap size on establishment, growth Plant Industry Bulletin, P476. and flowering of the endangered Campbell, M.H. (1982). The biology of Rutidosis leptorrhynchoides (Asteraceae). Australian weeds 9. Nassella trichotoma Journal of Applied Ecology 34, 566-76. (Nees) Arech. The Journal of the Osbourne, W., Kukolic, K. and Jones, S. Australian Institute of Agricultural Sci- (1993). Management of threatened ver- ence 48, 76-84. tebrates in native grasslands: a case of Carr, G.W., Yugovic, J.V. and Robinson, grasping at straws. In ‘Management of K.E. (1992). ‘Environmental weed inva- relict grasslands’ Proceedings of a sions in Victoria: conservation and im- workshop and public seminar, ACT plications for management’. (Depart- Parks and Conservation Service, Can- ment of Conservation and Environ- berra, pp. 89-95. ment, Melbourne). Phillips, A. and Hocking, C. (1996). A Connor, H.E. (1987). Reproductive biol- method of replacing serrated tussock ogy in the grasses. In ‘Grass systemat- with weed-free kangaroo grass in de- ics and evolution’, eds. T.R. graded native grassland remnants. Pro- Soderstrom, K.W. Hilu, C.S. Campbell ceedings of the Eleventh Australian and M.E. Barkworth, pp. 117-32. Weeds Conference, pp. 524. (Victorian (Smithsonian Institute Press, Washing- Weed Science Society, Melbourne). ton, USA). Rosengurtt, B. (1984). Gramineas cleisto-