The role of biological control agents in an IWM program for Chrysanthemoides monilifera subsp. rotundata (bitou bush)

Royce H. Holtkamp1

Summary Bitou bush, Chrysanthemoides monilifera subspecies rotundata, is a native of South Africa, which was used extensively in Australia as a sand-stabilising plant and for revegetation of coastal areas mined for mineral sands. It has now become a serious environmental weed in eastern Australia, primarily of conservation areas, where it significantly reduces biodiversity. Since 1989, six species of insects have been released on bitou bush, four of which have established. These are having varied impacts on bitou bush with bitou tip moth, Comostolopsis germana, and bitou seed fly, Mesoclanis polana, being the most successful. An integrated weed management approach appears to be the best option for long-term sustainable control of bitou bush. This paper discusses the use of biological control agents in combina- tion with other control options such as strategic herbicide applications, fire, physical removal and revegetation techniques.

Keywords: biological control, bitou bush, Chrysanthemoides monilifera subspecies rotundata, integrated weed management.

The plant withdrawn. However, this action came far too late and by 1976 C. monilifera subsp. rotundata was naturalised Chrysanthemoides monilifera subspecies rotundata along much of the NSW coast. (DC.) T. Norl. (bitou bush), is a competitive environ- A survey conducted in 2001 by the NSW National mental weed of South African origin. It is primarily Parks and Wildlife Service (NPWS) has shown C. restricted to areas of summer rainfall (Parsons and monilifera subsp. rotundata to be present on 900 km Cuthbertson 1992) and infests coastal areas of southern (80%) of the NSW coastline and the dominant plant on Queensland, New South Wales (NSW) and Lord Howe over 400 km. Over approximately two-thirds of this Island. There is also a localised inland infestation at area, it could completely dominate and eventually Menindee Lakes, NSW. In NSW it is common north of displace most of the existing native vegetation. This Sydney and occurs south to the Victorian border. current distribution represents a 36% increase in the C. monilifera subsp. rotundata was first recorded in area over which it was present in a 1982 survey that was Australia from Stockton near Newcastle in 1908 (Weiss also conducted by NPWS (Holtkamp et al. 1999). et al. 1998) where it appears to have been an accidental The importance of C. monilifera subsp. rotundata introduction in ships ballast. From 1946 to 1968, was officially recognized in early 1999 by both the C. monilifera subsp. rotundata was used as a sand- NSW National Parks and Wildlife Service which listed stabilising plant and to revegetate coastal areas mined it as a “key threatening process” under the Threatened for mineral sands. The capacity of C. monilifera subsp. Species Conservation Act 1995 and by the Common- rotundata to invade native vegetation was then recog- wealth of Australia which listed it as a Weed of nised and its recommendation for coastal planting was National Significance under the National Weeds Strategy 1997.

1 C. monilifera subsp. rotundata is largely an environ- NSW Agriculture, Weed Biological Control Unit and Cooperative Research Centre for Australian Weed Management, RMB 944, Calala mental weed as it is easily controlled by stock grazing Lane, Tamworth, NSW 2340, Australia. and cultivation. It is primarily restricted to non-agricul-

412 Biological control in IWM of bitou bush tural areas such as national parks, forests, coastal dune drought this species proved extremely difficult to estab- ecosystems and other recreational land. The impact and lish. Multiple releases were made at sites that had the control of C. monilifera subsp. rotundata have been best plant quality and there are signs of a Tortrix sp. discussed in more detail by Holtkamp et al. (1999) and population persisting at some of these sites. It now Vranjic (2000). appears as if the drought is breaking and high hopes are held for this insect. Biological control Integrated weed management A biological control program against C. monilifera (which includes C. monilifera subsp. rotundata and Despite the success of two biological control agents and C. monilifera subsp. monilifera (boneseed)) was expectations of success for another, it is apparent that approved by Standing Committee on Agriculture in the only viable answer for long-term control of C. 1987. Surveys in South Africa have indicated that there monilifera subsp. rotundata is the integrated weed are more than 100 species of phytophagous insects management (IWM) approach discussed by Vranjic associated with the Chrysanthemoides species complex (2000). This includes such strategies as biological (Scott & Adair 1990). Eighteen of these species were control, physical removal, herbicides and fire. In the identified as having potential for the biological control past, these traditional techniques have been used to of C. monilifera. reduce infestations and limit spread of C. monilifera Six of these species have now been released on C. subsp. rotundata. Unfortunately, these techniques are monilifera subsp. rotundata, but only four have success- limited in their use for C. monilifera subsp. rotundata fully established. These are: bitou tip moth, Comostol- control for a number of reasons. opsis germana Prout; bitou tortoise beetle, Cassida sp.; Physical control techniques are extremely labour bitou seed fly, Mesoclanis polana Munro; and bitou intensive and are usually carried out by local volunteer leaf roller, Tortrix sp. groups targetting small areas. These groups mainly More than 200 releases of C. germana were made organise working parties to remove C. monilifera subsp. between 1990 and 1997 at 72 sites in NSW, ranging rotundata plants by hand pulling, although painting cut from the Queensland border to Tathra in southern stumps with glyphosate is also practised. The cut-stump NSW. It is now established along most of the NSW method is preferred by many workers because it results coast and it is believed that this insect has spread in minimal soil disturbance and subsequent erosion. throughout all C. monilifera subsp. rotundata infesta- Physical removal is particularly effective in small areas tions, with population levels still increasing in some of high conservation significance. Larger scale control areas. Populations in excess of 400 larvae m–2 have using these methods is not practical because it is too occurred at some sites (Holtkamp unpublished data) labour intensive. The possibility of removing C. monil- despite the presence of two hymenopteran parasitoids, ifera subsp. rotundata in areas infested for many years one of which parasitizes up to 50% of C. germana is compounded by large soil seed banks. Weiss and larvae (Holtkamp 1993). In many areas C. germana is Milton (1984) recorded a soil seed bank of 2030 seeds having a significant impact on flowering and seed per m2 near Moruya on the south coast of NSW, and production of C. monilifera subsp. rotundata. Holtkamp (unpublished data) has recorded a soil seed M. polana was first released in very low numbers in bank of up to 1968 viable seeds per m2 at Port August 1996 at Iluka Bluff and Dunbogan. Since then, Macquarie. nine releases have been made on the NSW North Coast. Herbicides aerially applied using helicopters have By August 1998, M. polana had been found from near proven to be extremely effective for broad scale C. Fraser Island in Queensland to Tathra, a total of over monilifera subsp. rotundata control (Toth et al. 1996). 1200 km of coastline (Edwards et al. 1999). Over much There is a “window of opportunity” during the winter of this area, population levels are extremely high and period immediately following peak flowering. At this reductions in seed production in excess of 50% have time C. monilifera subsp. rotundata plants are highly been recorded. susceptible to the herbicides glyphosate and metsul- Cassida sp. was first released at La Perouse (a furon methyl, while over 180 native species tested were suburb of Sydney) in 1995. A total of 12 releases was virtually unaffected. Further herbicide treatments are made, with locations spread over most of the NSW required approximately every two years until the soil coast. Recent surveys have shown it to be present at all seed bank is exhausted. It is important that none of the of these sites, but only close to the original release sites regenerating plants be allowed to flower and set seed. and only in low numbers. It would seem that the likely Unfortunately, herbicide application is not suitable in a impact of Cassida sp. will be minimal. number of situations such as in the presence of rare or Releases of Tortrix sp. commenced in 2001. There threatened flora. have now been more than 50 releases made at 20 sites The interactions between biological control agents along the NSW coast in a combined CSIRO Ento- and herbicides were discussed by Ainsworth and mology/NSW Agriculture project. Due to prolonged Holtkamp (1999), who reached the conclusion that

413 Proceedings of the XI International Symposium on Biological Control of Weeds herbicide application was unlikely to significantly References affect populations of M. polana. However, any inte- grated program which incorporates herbicides and Ainsworth, N. and Holtkamp, R.H. (1999) Integrating herbicide biological control will need to consider all species of use with biological control of bitou bush. Proceedings of the biological control agents present at the time of treat- 12th Australian Weeds Conference (eds A.C. Bishop, M. ment, to ensure that sufficient agents remain to allow Boersma & C.D. Barnes), pp. 275–278. Tasmanian Weed reestablishment. Society, Hobart, Australia. The use of fire, especially following herbicide appli- Edwards, P.B., Holtkamp, R.H. and Adair, R.J (1999) Estab- cation, stimulates germination of virtually the entire C. lishment and rapid spread of the bitou seed fly, Mesoclanis polana Munro (Diptera: Tephritidae), in eastern Australia. monilifera subsp. rotundata soil seed bank. This then Australian Journal of Entomology 38, 148–150. leaves these seedlings vulnerable to attack by foliage feeding biological control agents or further herbicide Holtkamp, R.H. (1993) Parasitoids of Comostolopsis germana (Lepidoptera: Geometridae). Journal of the Australian Ento- applications. However, fire is not suitable for all areas mological Society 32, 387–388. because ecosystems such as coastal dunes and rainfor- Holtkamp, R.H., Edwards, P.B. and Adair, R.J. (1999) Biolog- ests are not fire adapted. ical control of bitou bush: just around the turn of the millen- Integrated control of C. monilifera subsp. rotundata nium? Proceedings of the 12th Australian Weeds has the potential to reduce this weed to a minor compo- Conference (eds A.C. Bishop, M. Boersma & C.D. Barnes), nent of invaded vegetation, but will never eradicate it. pp. 667–669. Tasmanian Weed Society, Hobart, Australia. Any integrated program will have to ensure that suffi- Parsons, W.T. and Cuthbertson, E.G. (1992) Noxious Weeds of cient biological control agents remain following other Australia. Inkata Press, Melbourne, Australia. forms of treatment to ensure reestablishment of biolog- Scott, J.K. and Adair, R.J. (1990) The commencement of ical control agent populations. Continuing physical and biological control of bitou bush and boneseed (Chrysanthe- herbicidal control by volunteer groups in areas of high moides monilifera). Proceedings of the 9th Australian conservation significance is also important. The regen- Weeds Conference (ed J.W. Heap), pp. 126–129. Crop eration of coastal areas cleared of C. monilifera subsp. Science Society of South Australia, Adelaide, Australia. rotundata by local volunteer groups also forms an Toth, J., Milham, P.J., Meszaros, I., Kaldor, C.J., Fullerton, important component of this program. It is essential that R.N. and Burrows, F. (1996) Research on chemical control revegetation of disturbed habitat occurs quickly, to of bitou bush in New South Wales. Proceedings of the 11th prevent the niche previously occupied by C. monilifera Australian Weeds Conference (ed R.C.H. Shepherd), pp. subsp. rotundata from being occupied by C. monilifera 468–475. Weed Science Society of Victoria, Melbourne, Australia. subsp. rotundata seedlings or by another, perhaps more serious, weed species. Vranjic, J. (2000) Best Practice Management Guide 3. Bitou bush, Chrysanthemoides monilifera subsp. rotundata. CRC for Weed Management Systems, Adelaide, Australia. Acknowledgements Weiss, P.W., Adair, R.J. & Edwards, P.B. (1998) Chrysanthe- moides monilifera (L.) T. Norl. in The Biology of Australian Research on biological control of C. monilifera was partly Weeds. Vol. 2 (eds. F.D. Panetta, R.H. Groves, and R.C.H. funded by the Australian and New Zealand Environment Shepherd) pp. 49–61. R.G. and F.J. Richardson, Melbourne, and Conservation Council (ANZECC) and administered Australia. by the ANZECC Task Force for Weeds of Conservation Weiss, P.W. and Milton, S.J. (1984) Chrysanthemoides monil- Concern and the CRC for Weed Management Systems. ifera and Acacia longifolia in Australia and South Africa. Additional funding was provided by the CRC for Weed Proceedings of the 4th International Conference on Medi- Management Systems and the Natural Heritage Trust. terranean Ecosystems. pp. 159–160. Perth, Australia.

414