94 Protection Quarterly Vol.13(2) 1998 itch grass (Rottboellia cochinchinensis) as a Biological control of serrated tussock ( case study, he found that highly specific fungal pathogens occur on this grassy trichotoma): Is it worth pursuing? weed in its centre of origin (Evans 1995). Indeed, both the rust and smut species D.T. BrieseA and H.C. EvansB under evaluation proved to be too specific A CSIRO Entomology, GPO Box 1700, Canberra, ACT 2601, Australia. with clearly defined pathotype-biotype associations. Thus, it is probable that a B IIBC, Silwood Park, Ascot, Berkshire, SL5 7TA, United Kingdom. suite of pathotypes will have to be selected in order to cover the range of weed bio- Summary measures, revolving around the use of the types present in the neotropical target area. The potential for biological control of the herbicide fluproponate, have not pre- Nonetheless, the view that natural en- weedy grass, Nassella trichotoma, in vented the continued spread of the weed emies of grasses are not sufficiently spe- Australia is discussed in the light of re- (Campbell and Vere 1995). It is clear that cific and that the risk to economically-im- cent surveys for pathogens in its native other control options need to be investi- portant species of grass is too great to per- range and an improved understanding of gated in order to develop a more effective mit implementation of classical biological the taxonomic relationships of Nassella management strategy for this weed. One control remains a common one, and has to other Australian and South American such option is classical biological control; discouraged such projects in Australia. genera of the tribe . This informa- the introduction of a highly specific natu- This paper therefore re-examines the po- tion gives increased optimism that a suit- ral enemy(ies) from their native range that tential for biological control of N. tricho- able biological control agent or agents is capable of killing N. trichotoma or toma, in the light of an improved under- can be found. The type of research re- reducing their vigour, capacity for disper- standing of its taxonomic affinities and re- quired to do this is briefly discussed. sal and ability to compete with native and cent preliminary surveys for pathogen pasture grasses. natural enemies in its native range. Introduction The first survey of N. trichotoma was Serrated tussock (Nassella trichotoma undertaken by Wells (1977) in Argentina, Agent host-range and the taxonomic (Nees) Arech.) is a perennial grass of to assess the potential for its biological relationship of Nassella to other southern South American origin that was control in South Africa. He found that it grasses accidentally introduced into Australia in was not an aggressive or weedy plant in Given Wapshere’s (1990) contention that the early 1900s and has since become a se- its native range, due in part to the effects N. trichotoma is ‘probably too closely re- rious weed problem in south-eastern Aus- of natural enemies, including insects, lated to native spp. to allow the in- tralia (Campbell and Vere 1995). It has fungi and rodents. A follow-up survey for troduction of agents’, it is pertinent to re- also become a weed in New Zealand, insect natural enemies was made by Erb examine the taxonomic position of ser- South Africa and the United States. Ser- (1988), that identified 10 species of weevil rated tussock. Its relationship to other rated tussock is highly invasive, produc- and two moths. Unfortunately, while some grass taxa has long been problematic, at ing large quantities of seed that are widely species specialized on tussock grasses, all times being placed within the large genus dispersed by wind and capable of forming insects attacked a range of other grass spe- Stipa containing several hundred species dense infestations. cies. Evans (1991) reviewed the literature in Eurasia, South America and Australia. In Australia, the most seriously affected on arthropod natural enemies of grassy The one point of agreement has been that country, infestations occur mainly on pas- weeds and concluded that there is an ab- serrated tussock belongs to the tribe ture land, where the weed greatly reduces sence of records of host specific insects Stipeae, though this tribe has variously carrying capacity as it is unpalatable to which may be due to the simplistic and been placed in the subfamilies Pooidae stock. If forced to graze it, sheep lose uniform architecture of the Graminae and Arundinoidae in different morpho- weight and may die (Campbell and Vere which favours polyphagy and reduces the logical classifications of the (Clay- 1995). Increasingly, serrated tussock is in- evolutionary pressures for monophagy. ton and Renvoize 1985, Watson et al. 1985). vading natural environments, in National While Erb (1988) confirmed that ‘the com- Recent molecular phylogenetic evi- Parks and water catchment areas, where it bination of insect and fungus attack can dence (Hsaio et al. 1995, Barker et al. 1995) threatens native grasslands and modifies provide a control’ of serrated tussock, suggests that Stipeae does not sit comfort- natural habitats. The most recent survey South Africa did not pursue this work. ably in either sub-family, though only of its distribution in New South Wales Wapshere (1990) examined the possi- Simon (1993) has erected a new subfamily (Jones and Vere 1998) found that serrated bilities of controlling Australian grassy (Stipoideae) for it. Simon (1993) listed only tussock occupies 880 000 ha, indicating weeds biologically. He reported that, al- four genera within the tribe Stipeae that continued spread in the past 20 years. though no fungus had been recorded from are present in Australia: Nassella contain- Moreover, an additional 2 million ha have N. trichotoma, several rust and smut fungi ing six introduced species (McLaren et al. occasional plants present and can be con- infected both members of the genus 1998), Stipa containing over 60 native spe- sidered at risk. In Victoria, infestations Nassella and the closely related Stipa in cies (Vickery et al. 1986), rep- have increased to 130 000 ha with a poten- South America. This led him to conclude resented by a single species, P. miliacea, tial range of 4 million ha (D. McLaren, per- that biological control was unlikely for introduced as a pasture grass but not cur- sonal communication). In Tasmania, the serrated tussock, because Stipa contains rently utilized widely, and , area infested is smaller (c. 1500 ha) but several native Australian species that are represented by a single weedy species in- spreading (Goninon 1998). The economic useful pasture grasses and the risk to them troduced from South America (Simon losses attributed to serrated tussock are would preclude the introduction of any 1993). currently of the order of $40m per year in pathogen. In contrast, Evans (1991) as- Nassella has at various times been con- New South Wales (Jones and Vere 1998) sessed the potential of fungal pathogens sidered a subgenus of Stipa or a genus in and $5m per year in Victoria (Nicholson as biological control agents of tropical its own right (Barkworth 1990), while N. 1997). grassy weeds and considered that, al- trichotoma has been shifted between the Serrated tussock is a declared noxious though there was still an urgent require- two genera. Nassella was first revised by weed in New South Wales, Victoria and ment for more basic research, the pros- Parodi (1947) who excluded N. trichotoma, Tasmania. Despite this, current control pects appeared to be promising. Using while Corti (1951) assigned it as the Plant Protection Quarterly Vol.13(2) 1998 95 unique member of the section Nassellopsis Eurasian lines within Stipa. This is the classification Piptatherum maintained today by some South Ameri- 1 2 3 can botanists (Zuloaga et al. 1994), Stipa Trikeraia whereas Australian taxonomists included it within the genus Nassella (Vickery and * Jacobs 1980). In the most recent revision, Obtusae * Barkworth (1990) redefined the circum- scription of the genus to include all spe- Boreobtusae cies of Stipeae with strongly overlapping lemma margins and lemma apices that are fused into a crown. This expanded genus Nassella comprises 79 species including N. tricho- 5 Stipa (sect. Stephanostipa) toma, N. neesiana and 60 South American 4 Austrostipa * ArchaeostipaArchaeostip a species formerly assigned to Stipa. 3 Piptochaetium 2 Barkworth and Everett’s (1987) exami- Podopogon nation of the relationships of taxa within 1 Hesperostipa * the Stipeae, including for the first time a comparison of Australian and non-Aus- Australian lines American lines tralian Stipa species, sheds most light on Figure. 1. The phylogenetic relationships of monophyletic taxa in the tribe taxonomic position vis a vis biological con- Stipeae. Numbers indicate where several monophyletic groups occur within trol. They identified a number of mono- phyletic species-groups, including five en- one genus, and asterisks indicate those genera or monophyletic groups demic to Australia, covering the previ- containing species previously assigned to the genus Stipa. There are several ously described genera (Figure 1). While other groups within the American lines of Stipa. Figure modified from South American species within Stipa were Barkworth and Everett (1987). close to species of Nassella, both these groups were quite distinct from the Aus- Table 1. Genera of the tribe Stipeae present in Australia that would need to tralian groups of Stipa (Figure 1). In fact, be considered in any host-testing program for biological control agents of these differences are considered sufficient serrated tussock. to warrant erecting a new genus, Austro- stipa comprising 13 subgenera, to describe Genus No. of species Origin Status the Australian native taxa (Jacobs and Austrostipa 62 Australia native taxon with several useful Everett 1996). In addition, all three exotic pasture and several weedy species species present in Australia that were as- Nassella 6 South America declared noxious weeds or signed to Stipa have recently been placed potentially weedy in other genera; S. brachychaeta and S. caudata are now considered to belong to Achnatherum 2 South America potentially serious weeds the Eurasian genus Achnatherum (Jacobs et 1 South America localized minor weed al. 1995), while the recent South American Piptochaetium 1 South America localized minor weed invader, formerly known as S. papposa, has Piptatherum 1 Eurasia minor pasture grass and used in been renamed Jarava plumosa (Jacobs and mine reclamation Everett 1997). The stipoid genera present in Australia, as currently interpreted, are would probably need to be conservative week’s duration was undertaken in March shown in Table 1. and include selected grasses from even- 1995 (Evans and Ellison 1995), followed Given the differences between Austral- more distantly related taxa to assure all up by observations made in conjunction ian and American lineages, the relation- parties with interests in economically im- with surveys on other weed species dur- ships amongst South American species of portant grasses that any introduction was ing November 1995 and April 1996 (Evans the tribe Stipeae and the host range of safe. The final test list would therefore be 1996). Collections were made at 17 sites in pathogens found on them provides little quite long and would undoubtedly need central and western Argentina in March evidence of any potential threat to Aus- to include cereal, pasture, turf and native 1995 and at one further site in north west- tralian species. Nonetheless, it is clear that species. ern Argentina in April 1997 (Figure 2). genera of the tribe Stipeae contain the Nine species of fungi were identified from most closely related grasses to the weedy Recent South American surveys serrated tussock plants, and these are Nassella species in Australia, and the first Biological control agents are, by defini- listed in Table 2. step in a biological control program will tion, highly adapted to their host plant, Based on the initial survey, Evans and be to test any candidate pathogens against and the chances of finding such an organ- Ellison (1995) concurred with the conclu- representatives of the subgenera of ism will be greatest near the centre of di- sion of Wells (1977) that, in non-agricul- Austrostipa, before proceeding with a versification of the target weed genus. The tural or natural ecosystems in Argentina, wider host-test list. Following the cen- distribution of Nassella species (Barkworth N. trichotoma appears to be part of the in- trifugal phylogenetic testing sequence of 1990), shows clearly that this genus is cen- digenous flora rather than an invader and Wapshere (1974), subsequent testing tred in temperate South America, and that fungal pathogens would seem to play might involve species of the four intro- Wells (1977) postulated that N. trichotoma a role in controlling Nassella populations. duced Stipeae genera represented in Aus- originated in the Sierra de Ventana region Some of these pathogens have very tralia (Table 1), followed by representa- of Argentina. Hence, in 1995 CSIRO Ento- broad host and geographic ranges. For ex- tives of other tribes in the Arundinoideae mology engaged IIBC to survey popula- ample, the smut Ustilago hypodytes is al- and . Given that there has been tions of N. trichotoma in Argentina for ready known from a range of Australian no previous introduction of a pathogen pathogens that may have potential as stipoid grasses. Others appear to be for the biological control of a grass, this list biocontrol agents. The main survey of one much more specific. The two pathogens of 96 Plant Protection Quarterly Vol.13(2) 1998 Table 2. Fungi collected by H.C. Evans and C.A. Ellison during surveys of Nassella trichotoma in Argentina in 1995 and 1996. Fungal pathogen No. of sites Symptoms (out of 18) Basidiomycota: Basidiomycetes: Stereales Corticium sp.A 11 Infected plants having snow-white mycelia at stem / root interface, turning to grey on centres of dead or dying tussocks Basidiomycota: Teliomycetes: Uredinales Puccinia nas(s)ellaeB Arth. and Holw. 2 Rust pustules within tightly rolled leaves on plants suffering dieback. More common on young plants at one site Uredo sp. 1 Basidiomycota: Ustomycetes: Ustilaginales Ustilago hypodytes (Schlecht.) Fr. 1 Inflorescences replaced by smutted heads Deuteromycotina: Coelomycetes Aschochyta leptospora var. variispora Punith. 2 Associated with leaf chlorosis and subsequent necrosis Stagonospora sp.A 7 Associated with leaf chlorosis and subsequent dieback of leaves and flowering stems Hendersonula sp.A 1 Septoria sp.A 1 Deuteromycotina: Agonomycetes Rhizoctonia sp. 1 Leaf bases colonized by network of dark brown hyphae associated with chlorosis A All probably represent undescribed species (fide IMI Reports). B Orthographic error; in this publication, and continued in subsequent ones (Greene and Cummins 1958), the host genus was listed as Nasella, hence the specific epithet nasellae. greatest immediate interest for biological than serrated tussock belonging to a large control are Corticium sp. and Puccinia cosmopolitan genus Stipa, there are sev- nas(s)ellae. Corticium sp. was the most com- eral genera involved, with Austrostipa mon fungus found during the field sur- present in Australia, and an enlarged veys, forming a mycelial mat in the base of Nassella, including N. trichotoma, N. the tussocks and being associated with se- neesiana and many South American spe- vere tussock decline at 11 of the 18 sites cies formerly in Stipa, centred in South sampled (Figure 2). Tussock grasses from America. The likely Gondwanan origins other species that occurred in the same of the Stipeae (Barkworth and Everett area were not affected with these symp- 1987) suggest a long period of separate toms, suggesting that there may be a high evolution of the taxa and their associated degree of specificity. There has been no mycofloras. The difference in mycofloras taxonomic treatment of this species out- of South American and Australian stipoid side Europe and and it re- taxa is indicated by the fact that, based on Buenos Aires mains undescribed. Rust fungi have al- collection material held in New South ready been successfully used in a number Wales and Victoria, none of the 27 species of biological control projects (Julien 1992), of fungi known from Austrostipa in Aus- and one such species, provisionally iden- tralia have been recorded on introduced tified as P. nas(s)ellae, was found heavily Nassella. Moreover, none of the rusts attacking the leaves of a number of ser- found on South American taxa are known rated tussock populations at one location from Australian ones, while Puccinia in northwestern Argentina (Figure 2). Ac- flavescentis, common on Austrostipa spp. is Figure 2. Collection sites of fungi cording to Greene and Cummins (1958) not recorded from South America found on Nassella trichotoma in and Cummins (1971), P. nas(s)ellae is (Cummins 1971). Argentina during 1995–96. Closed known only from other species of Nassella In addition, revision of the stipoid taxa circles = sites with heavy attack by (following Barkworth’s (1990) revision of suggests that some pathogens from Corticium sp. Open squares = sites the genus). However, the form collected Nassella may be even more specific than with heavy attack by Puccinia sp. has smaller urediniospores (16–18 × 19–20 previously thought, as none of the rusts mm) than P. nas(s)ellae sensu stricto (23–26 described by Cummins (1971) from var. platensis. Surprisingly, N. trichotoma × 26–30 mm according to Cummins 1971) Nassella spp. had recorded host ranges was not included as a host of this, or any and could well be an undescribed taxon outside the tribe Stipeae, and five of the other rust species, in Argentina (Lindquist (R. Treu personal communication). Aecia nine species only had hosts within the ge- 1982). are unknown (Green and Cummins 1958, nus itself. Indeed, there is now taxonomic In fact, prior to the current surveys, no Lindquist 1982), and thus it has not been evidence for the existence of host specific pathogens had previously been recorded proven if the rust is autoecious or heteroe- forms (races, strains, varieties) within from N. trichotoma. If nine species were re- cious. This can only be established by de- Puccinia nas(s)ellae. Lindquist (1982), in his corded during such a brief investigation, tailed field studies. monograph of Argentinian rusts, listed it is highly likely that further fungal several hosts for P. nas(s)ellae, sensu pathogens would be collected should Discussion Arthur and Holway (Greene and more widespread surveys over a full The most recent taxonomic studies have Cummins 1958): N. caespitosa, N. chilensis range of seasons be possible. The fact that highlighted the polyphyletic origins of the and ‘Stipa’ sp., but erected a new variety the Australian and American taxa are less genus Stipa. They indicate that, rather for the rust on ‘Stipa’ neesiana, P. nas(s)ellae closely related than previously thought, Plant Protection Quarterly Vol.13(2) 1998 97 the indication that these grasses harbour a References names for Australasian species for- rich and incompletely known mycoflora Barker, N.P., Linder, H.P. and Harley, merly included in Stipa (Gramineae). and the apparently high degree of E.H. (1995). Polyphyly of Arun- Telopea 6, 579-95. specificity in the known range of some dinoideae (Poaceae): evidence from Jacobs, S.W.L. and Everett, J. (1997). Jarava pathogens collected from them gives rbcL sequence data. Systematic Botany plumosa (Gramineae), a new combina- greater optimism for finding a suitable 20, 423-35. tion for the species formerly known as biological control agent. While accepting Barkworth, M.E. (1990). Nassella Stipa papposa. Telopea 7, 301-2. that any biological control project carries a (Gramineae: Stipeae): Revised interpre- Jacobs, S.W.L., Everett, J. and Barkworth, risk of failure, there is ample justification tation and nomenclatural changes. M.E. (1995). Clarification of morpho- to pursue the detailed studies needed in Taxon 39, 597-614. logical terms used in the Stipeae South America to accurately determine Barkworth, M.E. and Everett, J. (1987). (Gramineae), and a reassessment of the potential of pathogens for control of Evolution in the Stipeae: Identification Nassella in Australia. Taxon 44, 33-41. serrated tussock. Such a project would and relationships of its monophyletic Jones, R.E. and Vere, D.T. (1998). The eco- also provide the opportunity to investi- taxa. In ‘Grass Systematics and Evolu- nomics of serrated tussock in New gate the potential of pathogens to control tion’, eds. T.R. Soderstrom, K.W. Hilu, South Wales. Plant Protection Quarterly the related Chilean needle grass, N. C.S. Campbell, and M.E. Barkworth, 13, 70-6. neesiana, an emerging weed problem in pp. 251-64. (Smithsonian Institution, Julien, M.H. (1992). ‘Biological control of south-eastern Australia. ). weeds: A world catalogue of agents These studies should aim to: Campbell, M.H. and Vere, D.T. (1995). and their target weeds. 3rd edition’, 186 • initially carry out detailed surveys Nassella trichotoma (Nees) Arech. In ‘The pp. (CABI, Wallingford, UK). throughout the range of N. trichotoma Biology of Australian Weeds Volume Lindquist, J.C. (1982). Royas de la and N. neesiana (concentrating on those 1’, eds. R.H. Groves, R.C.H. Shepherd Républica Argentina y zonas limitrofes. areas that are ecoclimatically most and R.G. Richardson, pp. 189-202. (R.G. INTA, Buenos Aires. similar to infested areas of Australia) and F.J. Richardson, Melbourne). McLaren, D.A., Stajsic, V. and Gardener, for pathogens with biological control Clayton, W.D. and Renvoize, S.A. (1986). M.R. (1998). The distribution and im- potential and select a short-list of ‘Genera Gramineum: Grasses of the pact of South/North American stipoid pathogens for more detailed study, World’. (HMSO, London). grasses (Poaceae: Stipeae) in Australia. • investigate the specificity of these path- Corti, R. (1951). Stipa trichotoma Nees nella Plant Protection Quarterly 13, 62-70. ogens, starting with Puccinia nas(s)ellae selvadi San Rossofe, nuovo inquilino Nicholson, C. (1997). The cost of serrated and Corticium sp., against key Austral- dell Flora Toscana. Instituto Botanico del tussock control in central western Vic- ian species such as Austrostipa to deter- Universiti de Firenze 36, 475-86. toria. Unpublished report of the Victo- mine whether it is safe to proceed with Cummins, G.B. (1971). ‘The rust fungi of rian serrated tussock working group. more detailed studies, and if so, cereals, grasses and bamboos’, 565 pp. Parodi, L.R. (1947). Las especies de • study their epidemiology and impact (Springer-Verlag, Berlin). gramíneas del género Nassella de la Ar- on the population dynamics of selected Erb, H.E. (1988) Stipa trichotoma. Some gentina y Chile. Darwinia 7, 369-95. field populations of N. trichotoma and considerations on the species, its habi- Simon, B.K. (1993). ‘A key to Australian N. neesiana, and concurrently tat and natural distribution in Argen- grasses. 2nd edition’, 206 pp. (Queens- • develop culture methods for them in tina. Unpublished report, 8 pp. land Department of Primary industries, the laboratory and study their biology, Evans, H.C. (1991). Biological control of Brisbane). life-cycle and virulence against Aus- tropical grassy weeds. In ‘Tropical Vickery, J.W. and Jacobs, S.W.L. (1980). tralian forms of the weeds. Grassy Weeds,’ eds. F.W.G. Baker and Nassella and Oryzopsis in New South The end-point would be the identification P.J. Terry, pp. 52-72. (CAB Interna- Wales. Telopea 2, 17-23. of one or more pathogens with demon- tional, Wallingford, Oxon). Vickery, J.W., Jacobs, S.W.L. and Everett, strated potential for biological control, a Evans, H.C. (1995). Fungi as biocontrol J. (1986). Taxonomic studies in Stipa well-understood biology and having agents of weeds: a tropical perspective. (Poaceae) in Australia. Telopea 3, 1-132. proven specificity against key grass spe- Canadian Journal of Botany 73, 58-64. Wapshere, A.J. (1974). A strategy for cies. The project would then be ready to Evans, H.C. (1996). Report of a survey for evaluating the safety of organisms for enter a second phase aimed at quarantine fungal pathogens of Nassella trichotoma biological weed control. Annals of Ap- clearance and eventual release as part of in Argentina, April 1996. Unpublished plied Biology 77, 201-11. the management strategy for weedy Report to CSIRO, IIBC, Ascot, 2 pp. Wapshere, A.J. (1990). Biological control Nassella species in Australia. Evans, H.C. and Ellison, C.A. (1995). Re- of grass weeds in Australia: an ap- port of a survey for fungal pathogens of praisal. Plant Protection Quarterly 5, Acknowledgments Nassella trichotoma in Argentina, March 62-75. We would like to thank Joy Everett, Sur- 1995. Unpublished Report to CSIRO, Watson, L., Clifford, H.T. and Dallwitz, rey Jacobs, David McLaren and Louise IIBC, Ascot, 13 pp. M.J. (1985). The classification of Morin for reviewing earlier versions of Greene, H.C. and Cummins, G.B. (1958). Poaceae: subfamilies and supertribes. this manuscript. Michael Priest (NSW Ag- A synopsis of the Uredinales which Australian Journal of Botany 33, 433-84. riculture) and Ian Pascoe (Agriculture Vic- parasitize grasses of the genera Stipa Wells, M.J. (1977). Progress with research toria) provided lists of pathogens col- and Nasella. Mycologia 50, 6-36. on Nassella tussock. Proceedings of the lected from Stipa (=Austrostipa) and Goninon, C. (1998). Serrated tussock in Second National Weeds Conference of Nassella in Australia. Bob Sproule and Tasmania. Plant Protection Quarterly 13, South Africa, Stellenbosch, Balkema, Gratton Wilson of the NSW Serrated Tus- 98-9. Cape Town, 48-55. sock Working Group organized the con- Hsaio, C., Chatterton, N.J., Asay, K.H. and Zuloaga, F.O., Nicora, E.G., Rugolo de sortium of funding to support the South Jensen, K.B. (1995). Molecular phylo- Agrasar, Z.E., Morrone, O., Pensiero, J. American surveys, which included the geny of the Pooideae (Poaceae) based and Cialdella, A.M. (1994) Catalogo de Cooma-Monaro Shire Council, NSW Na- on nuclear rDNA (ITS) sequences. Theo- la Familia Poaceae en la Republica Ar- tional Parks and Wildlife Service, NSW retical and Applied Genetics 90, 389-98. gentina. Monographs in Systematic Noxious Weeds Advisory Council, and Jacobs, S.W.L. and Everett, J. (1996). Botany from the Missouri Botanical Gar- the Snowy Mountains Authority. Austrostipa, a new genus, and new den 47, 1-178.