26 Protection Quarterly Vol.18(1) 2003 three treatments) with five replicates of each treatment. In August 2000, before Sparaxis bulbifera (Iridaceae) invading a clay based treatment, the number of S. bulbifera adults wetland on the – control methods (flowering ) and juveniles (non flow- ering plants) in each plot was counted and and observations on the reproductive biology a cover estimate (Braun-Blanquet) scored for each species (native and introduced) in Kate Brown and Kris Brooks, Environmental Weeds Action Network, each plot. In early October 2000 the her- c/- Swan Catchment Centre, PO Box 1906, Midland, 6936, bicide was carefully spot sprayed in each Australia. Email: [email protected] treatment. As this weed grows very closely amongst the native herbs and shrubs these were also subject to herbicide application. Metsulfuron methyl (Brushoff®, 600 g kg-1 metsulfuron methyl) was applied from a Summary competition from weeds such as Sparaxis 15 litre backpack sprayer at 2.5 g ha-1 and Sparaxis bulbifera, a cormous geophyte bulbifera. This South African cormous spe- at 1.0 g ha-1 and chlorsulfuron (Glean®, 750 from the Cape Region of South Africa, cies is a serious invader of clay based wet- g kg-1 chlorsulfuron) was applied at 2.5 g is a serious invasive weed of clay based lands on the Swan Coastal Plain (Hussey ha-1. All were applied with the penetrant wetlands on the Swan Coastal Plain. et al. 1997). Once established it forms dense Pulse® (1000 g L-1 polydimethylsiloxane) Where it invades these wetlands it forms monocultures displacing herbaceous flora at 2 mL L-1. The plots were rescored in dense monocultures, displacing much of in particular. Given that around 50% of the September 2001. Differences in counts be- the rich native herbaceous flora. Herbi- native flora at Brixton Street comprises an- tween treatments and years were analysed cide trials were conducted in the Brixton nual or perennial herbs, many of them rare using Kruskal Wallace ANOVA with post Street Wetlands south east of Perth, or restricted taxa (Keighery and Keighery hoc Mann-Whitney U-tests used to de- where it is invading shrublands and 1995), S. bulbifera poses a significant threat termine significance between individual herblands, and impacts of treatments to conservation values of the wetland. treatments. Plant nomenclature gener- on co-occurring native species recorded. Options for control of a weed growing ally follows Paczkowska and Chapman Metsulfuron methyl was trialed at 1.0 g closely among native plants in a wetland (2000). ha-1 and at 2.5 g ha-1 and chlorsulfuron situation are limited. At Brixton Street any Plants of S. bulbifera were also moni- at 2.5 g ha-1. All treatments significantly control program needs to take place over tored over time to determine bulbil and reduced the number of juvenile and a short time at the beginning of October as seed production and methods of spread. adult plants of S. bulbifera. Metsulfuron the wetland begins to dry out and before methyl at 2.5 g ha-1 was the most effective S. bulbifera comes into full flower. Hand Results reducing the number of adult plants per removal at this time is one option and is Sparaxis bulbifera control plot from a mean of 25.4 to 0.4 and the a particularly useful tool for removing All herbicide treatments significantly number of juvenile plants from a mean small isolated populations growing in reduced the number of both adult and ju- of 373.2 to 28.8. relatively undisturbed bushland. How- venile S. bulbifera plants in the trial plots. There were 27 species of native taxa ever hand removal is labour intensive and Metsulfuron methyl at 2.5 g ha-1 was the scattered across the trial plots and only consequently needs to be integrated with most effective followed by chlorsulfuron two taxa, Sowerbaea laxiflora and Phily- a carefully targeted chemical control pro- at 2.5 g ha-1 then metsulfuron methyl at drella drummondii appeared to have been gram. Trials in recent years have indicated 1.0 g ha-1. Metsulfuron methyl at 2.5 g adversely affected by the herbicide treat- that the sulfonylurea group of herbicides ha-1 was significantly more effective than ments. The results however were com- can be quite effective against a number of metsulfuron methyl at 1.0 g ha-1 in control- pounded by the spatial distribution and introduced bulbous and cormous species ling adult plants (Figure 1). Apart from the seasonal nature of the native flora. (Pritchard 1991, Dixon 1996, Pritchard this there were no significant differences Additional information on the reproduc- 1996, Peirce 1998, Meney 1999, Brown et between treatments. Although all treat- tive biology and spread of the weed was al. 2002). Observations and data indicate ments significantly reduced the number of also recorded. that some introduced bulbous or cormous juvenile plants, a mean of 28.8 (±15.08) in- species invading bushland can be control- dividuals per plot still remained following Introduction led with sulfonylurea herbicides with the most effective treatment (Figure 2). The Brixton Street Wetlands lie 20 km minimal impact on the native flora (Dixon south east of Perth at the foot of the 1996, Meney 1999, Moore 1999). Brown et Herbicide impacts on native plants Darling Scarp. A small 19 ha remnant al. (2002) in detailed studies on the con- There were 27 species of native plants on the winter wet flats of Guilford trol of Lachenalia reflexa with metsulfuron recorded from across the trial site; five formation clays, the wetlands have an methyl at 5 g ha-1 in a Banksia woodland shrubs, 14 perennial herbs (seven of those exceedingly diverse flora of 307 native found indications of minimum impact on geophytes) and eight annual herbs. Not taxa (Keighery and Keighery 1995). Spe- co-occurring native species. all species occurred in all plots. Of the five cies rich herblands cover the winter-wet With little published information shrubs, three species remained unchanged claypans, herb rich shrublands the clay on herbicide control of S. bulbifera this or increased in cover across all the herbi- flats and Eucalyptus calophylla woodland study aimed to trial the effectiveness of cide treatment plots they were present in. the higher ground where the soil is well two herbicides, metsulfuron methyl and Seedling recruits of Viminaria juncea and drained (Keighery and Keighery 1991). chlorsulfuron on S. bulbifera and to inves- Pimelia imbricata var. major were observed With this kind of habitat almost entirely tigate the impacts of these herbicides on in some treatment plots. Of the remaining cleared on the Swan Coastal Plain the co-occurring native species in the Brixton two taxa, Verticordia densiflora fluctuated area is of outstanding conservation value Street Wetlands. in cover across the chlorsulfuron plots (Keighery and Keighery 1991, Gibson while micrantha subsp. micrantha et al. 1994). One of the major threats to Methods decreased in cover, across both metsul- the native flora and to the plant com- Trial plots (1 m × 1 m) were laid out in furon treatments and fluctuated across munities of the wetlands is invasion and a randomized block design (control and the controls. Plant Protection Quarterly Vol.18(1) 2003 27 30 in the only treatment plots in which it occurred, metsulfuron methyl (1.0 g ha-1) (Table 1). 25 Observations on reproductive biology 20 and spread Observations over the period of our study have provided insights into some aspects 15 of the reproductive biology of S. bulbifera including seed biology, mechanisms of 10 dispersal and patterns of invasion within Mean number of plants the wetland. Although plants produce small bulbils up the stems as they die 5 down at the end of each season, popula- tions appear to spread mainly by seed. Each plant produces around 75 soft thin 0 papery coated seeds in late spring and control metsulfuron methyl metsulfuron methyl chlorsulfuron within the wetlands at least, water ap- 1.0 g per ha 2.5 g per ha pears to be a major dispersal agent. The seed has been observed floating in stand- Figure 1. Mean number (with standard error bars) of mature Sparaxis ing and flowing water and maps reveal bulbifera plants in treatment plots (1 × 1 m) before (dark) and after (pale) populations moving mainly through low herbicide treatments. lying wet areas including creeks and drains. Sheet water flow occurring across the wetland in winter may well explain 600 small populations of S. bulbifera appear- ing in undisturbed bushland. There is also evidence that human activities are playing 500 a role in the spread of the population with dense infestations occurring along walk- ing tracks. 400 Interestingly the seed appears to be relatively short lived in the soil, gener- ally germinating the first season following 300 ripening. Prolific seedling recruitment was observed throughout populations of S. 200 bulbifera before the management program Mean number of plants of number Mean began. One year following removal of all flowering plants from isolated popula- 100 tions little or no seedling recruitment was observed in those populations.

0 Discussion control metsulfuron methyl metsulfuron methyl chlorsulfuron Effectiveness of treatments and impacts 1.0 g per ha 2.5 g per ha on native taxa It is clear that the herbicides tested all ef- Figure 2. Mean number (with standard error bars) of juvenile Sparaxis fectively controlled S. bulbifera at the trial bulbifera plants in treatment plots (1 × 1 m) before (dark) and after (pale) sites, with metsulfuron methyl at 2.5 g -1 herbicide treatments. ha the most effective. The high number of juvenile plants remaining in the treat- ment plots indicates that there will need to be follow up control and any successful Of the 14 perennial herbs present, eight across all plots it occurred in including management strategy will require a com- taxa retained or increased cover within all controls. Philydrella drummondii decreased mitment of on-ground resources over a treatment plots in which they occurred in all the metsulfuron methyl (2.5 g ha-1) number of years. (Table 1). Cover fluctuated across one or and chlorsulfuron (2.5 g ha-1) treatment As not all native taxa occurred in all more herbicide treatment plots for two plots in which it occurred (Table 1). treatment or control plots, impacts of the taxa, Borya scirpoidea and Drosera menziesii There was a high level of recruitment three herbicide treatments on the native subsp. menziesii. Sowerbaea laxiflora de- into both control and herbicide treatment flora at Brixton Street is unclear. This is creased in the control plots it occurred in plots of annual herbs with four taxa retain- compounded by the highly seasonal na- and cover class fluctuated under both met- ing cover or recruiting into the treatment ture of the wetland flora. Nevertheless the sulfuron methyl at 2.5 g ha-1 and chlorsul- and control plots over 2000/2001 (Table results do indicate that effective control of furon at 2.5 g ha-1. In one of the chlorsul- 1). One taxon, Hydrocotyle alata, recruited S. bulbifera is possible with minimum im- furon treatment plots cover of Sowerbaea into control and metsulfuron (1.0 g ha-1) pacts on a range of native taxa. The only laxiflora decreased from 25–50% cover to treatments with cover fluctuating under shrub at the site that gave an indication zero. For one taxon, Cyathochaeta avenacea, the other two treatments. Of the remain- of being affected by metsulfuron treat- cover fluctuated across control plots only. ing three taxa, two retained cover in all ments, Kunzea micrantha subsp micrantha, Of the remaining two perennial herbs, treatment plots they occurred in and one, also showed a fluctuation in cover across Utricularia menziesii decreased in cover Drosera glanduligera, decreased in cover the control plots, increasing in some and 28 Plant Protection Quarterly Vol.18(1) 2003

Table 1. Changes in cover class among co-occurring native taxa following the three herbicide treatments to Sparaxis bulbifera. Taxa Life form Control Metsulfuron Metsulfuron Chlorsulfuron methyl methyl 2.5 g ha-1 1.0 g ha-1 2.5 g ha-1 Acacia lasiocarpa var. bracteolata shrub ≥ – ≥ ≥ Aphelia cyperoides annual herb – – ≥ ≥ Borya scirpoidea perennial herb ≥ ± ≥ – Burchardia congesta perennial herb-geophyte ≥ – – – Burchardia multiflora perennial herb-geophyte – ≥ – – Centrolepis aristata annual herb ≥ ≥ ≥ ≥ Chorizandra enodis perennial herb ≥ ≥ ≥ ≥ Chorizandra multiarticulata perennial herb – ≥ r r Cyathochaeta avenacea perennial herb ± ≥ ≥ – Drosera glanduligera annual herb – < – – Drosera menziesii subsp. menziesii perennial herb- geophyte ≥ ± ≥ r Goodenia micrantha annual herb r r r ≥ Hydrocotyle alata annual herb r r ± ± Isolepis cernua annual herb r r r r Kunzea micrantha subsp. micrantha shrub ± < < – Lomandra spp. perennial herb – ≥ – – Meeboldina canus perennial herb – ≥ – ~ Mesomelaena tetragona perennial herb – ≥ – – Pimelea imbricata var. major shrub ≥ – r ≥ Philydrella drummondii perennial herb-geophyte ≥ r < < Schoenus odontocarpus annual herb r r r r Siloxerus humifusus annual herb – r – r Sowerbaea laxiflora perennial herb-geophyte < – ± ± Tribonanthes longipetala perennial herb-geophyte r – r r Utricularia violacea perennial herb-geophyte < < < < Verticordia densiflora shrub ≥ ≥ ≥ ± Viminaria juncea shrub r r r – – not present in any plots before or after treatment, ≥ maintained or increased cover class in all plots it was present in, ± decreased in cover class in some and increased in others following treatment, < decreased cover class in all plots it was present in, r new recruit in plots in 2001.

decreasing in others, indicating that fac- requires further investigation. At the same time a spray contractor with a tors apart from herbicide application were Although the data is limited and dif- knowledge of the flora and a background impacting on the growth of the plants. ficult to interpret due to seasonal effects, in bushland work has carried out chemical With annual taxa in particular, it is dif- there are indications that S. bulbifera can control on heavier infestations of S. bulbif- ficult to determine if change in cover over be effectively controlled without impact- era in more disturbed areas. two years is seasonally related or treat- ing on a range of co-occurring native Effectiveness of the project over the ment related. Of the eight annual species taxa present at the study site. Given the three years has been reliant on workers present all but one maintained or increased high flora conservation values of these having an understanding of the distribu- in cover and this at least indicated that the wetlands this is an area worthy of further tion of S. bulbifera across the Brixton Street herbicide treatments were not adversely investigation. Wetlands. The populations were mapped affecting those species. Changes in cover at the start of the project, in 1998. The of perennial herbs, geophytes in particu- An integrated approach maps helped set priorities and allowed lar could also be seasonally related. Utri- Importantly, chemical control techniques for a carefully targeted works program cularia violacea, for example, disappeared are only one part of an effective strategy revisiting small isolated populations over from all plots it occurred in including con- for managing the spread of weeds such a number of years. The maps also allowed trols. Trials were scored almost a month as S. bulbifera into relatively undisturbed accurately recorded work to be carried out later in 2001 and this wetland perennial bushland. For the last three years bush re- over time and provided some information geophyte had probably died back to its generators have been employed through on the effectiveness of the management bulb. However the geophytes Philydrella September/October in the wetlands to strategy. drummondii and Sowerbaea laxiflora both manually remove small isolated popula- A detailed understanding of the repro- appear to have been adversely affected tions in undisturbed areas. The wetland ductive biology and the mechanisms for by the metsulfuron methyl (2.5 g ha-1) soils are still soft in early spring and entire dispersal of S. bulbifera is also an impor- and chlorsulfuron treatments and the im- plants including corms come out quickly tant part of any effective management pacts of these herbicides on these species and easily with minimal soil disturbance. strategy. The observation that most seed Plant Protection Quarterly Vol.18(1) 2003 29 of S. bulbifera germinates the season after References D.M. Richardson, F.J. Kruger, and H.J. ripening and is relatively short lived in the Braun-Blanquet, J. (1965). ‘Plant sociology: van Hensbergen). Ecological Studies 93, soil imply that once the adult population the study of plant communities’. (Haf- 145-60. (Springer Verlag, Berlin). has been controlled follow up work could ner, London). Meney, K. (1999). Mt. Eliza escarpment be over in a three to five years. Workers in Brown, K., Brooks, K., Madden, S. and restoration plan. (Kings Park and Bo- South Africa have found that among natu- Marshall, J. (2002). Control of the exotic tanic Garden, West Perth). ral populations with in the Cape Region of bulb, yellow soldier (Lachenalia reflexa) Moore, J. (1999). The tolerance of direct South Africa, the seeds of most geophytes invading a Banksia woodland, Perth, seeded native species to herbicides. do not display dormancy, generally ger- Western Australia. Ecological Manage- Proceedings of the 12th Australian minating the season after ripening (Keely ment and Restoration 3(1), 26-34. Weeds Conference, Hobart, eds A.C. and Bond 1997). Bond, W.J. and van Wilgen, B.W. (1996). Bishop, M. Boersma and C.D. Barnes, Although observations indicate that ‘Fire and plants’. (Chapman and Hall, pp. 529-34. (Tasmanian Weed Society water flow plays a major role in dispersal, London). Inc., Hobart). fire also probably facilitates the spread of Dixon, I.R. (1996). Control of bridal Paczkowska, G. and Chapman, A.R. S. bulbifera in the wetland. Bond and van creeper (Asparagus asparagoides) and (2000). ‘The Western Australian flora: Wilgen (1996) noted that South African distribution of Asparagus declinatus in a descriptive catalogue’. (Wildflower geophytes that occurred in the fire prone Kings Park bushland, 1991–1995. Pro- Society Inc., Western Australian Her- environment of the Cape Region of South ceedings of the bridal creeper sympo- barium and Botanic Gardens and Parks Africa were well adapted to surviving sium, eds J.P. Pigott, D.L. Lamont and Authority, Perth). summer fires. Soil offers very effective G.J. Keighery. Plant Protection Quarterly Pearce, G.A. (1963). Factors affecting the insulation to the summer dormant corms 11, 59-63. control of one leafed cape tulip. Thesis and some have contractile roots that pull Gibson, N., Keighery, B.J., Keighery, G.J., M.Sc. Agric., University of Western the storage organs of young plants deeper Burbidge, A.H. and Lyons M.N. (1994). Australia, 120 pp. into the soil helping them avoid lethal A floristic survey of the Southern Swan Peirce J.R. (1998). Oxalis pes-caprae L. In fire temperatures. Fire can break down Coastal Plain. Unpublished report for ‘The biology of Australian weeds’, eds bulb dormancy in some South African the Australian Heritage Commission F.D. Panetta, R.H. Groves and R.C.H. geophytes (Pearce 1963) and can stimu- prepared by the department of Con- Shepherd, pp. 141-56. (R.G. and F.J. late flowering followed by prolific seed servation and Land Management and Richardson, Melbourne). production in others (Goldblatt 1978, the Conservation Council of Western Pritchard, G.H. (1991). Control of bridal Richardson 1984, Le Maitre and Brown Australia (Inc.), Perth. creeper with herbicides. Plant Protec- 1992). The post fire environment then Goldblatt, P. (1978). An analysis of the flo- tion Quarterly 6, 126. offers space and light where seeds can ra of southern Africa; its characteristics, Pritchard, G.H. (1996). Efficacy of herbi- germinate and seedlings can establish relationships and origins. Annals of the cides against angled onion in pot tri- in the absence of larger trees and shrubs Missouri Botanic Garden 65, 369-436. als. Proceedings of the 11th Australian (Goldblatt and Manning 1998). Keeping Goldblatt, P. and Manning, J. (1998). Weeds Conference, Melbourne, ed. frequent fire out of the wetlands is prob- ‘Gladiolus in southern Africa’. (Fern- R.C.H. Shepherd, pp. 480-4. (Weed ably quite important in preventing further wood Press, Vlaeberg). Science Society of Victoria Inc., Mel- spread of S. bulbifera. Hussey, B.M.J., Keighery, G.J., Cousens, bourne). Weed control is a part of an overall site R.D., Dodd, J. and Lloyd, S.G. (1997). Richardson, G.R., Lubke, R.A. and Jacot restoration program at Brixton Street. The ‘Western weeds. A guide to the weeds Guillarmod, A. (1984). Regeneration protection of native plant communities is of Western Australia’. (The Plant Pro- of grassy fynbos near Grahamstown the focus rather than simple elimination tection Society of Western Australia, (eastern cape) after fire. South African of the weed. Species moving into the gap Victoria Park, Western Australia). Journal of 3, 153-62. created as S. bulbifera is removed will need Keeley, J.E. and Bond, W.J. (1997). Conver- to be carefully monitored. gent seed germination in South African fynbos and Californian chaparral. Plant Acknowledgements. Ecology 133, 153-67. Meg Brooks, Elizabeth Buters, Nick But- Keighery, G.J. and Keighery, B.J. (1991). ers, Kathryn Clarkson, Regina Drum- Floristics of reserves and bushland are- mond, Sally Madden, Sandra Sandich, as of the Perth Region (System Six). Part Leslie Shaw and Hayley Turner all spent II. Brixton Street Wetlands, Kenwick. In time at Brixton Street removing S. bulbif- ‘Floristics of reserves and bushland ar- era by hand. Bob Dixon and John Moore eas of the Perth Region (System 6), Parts provided valuable technical advice on II–IV’. (Wildflower Society of Western herbicide treatments. The Perth Branch Australia (Inc.), Nedlands). of the Wildflower Society (Inc.) provided Keighery, G. and Keighery, B. (1995). Flora. funding for the control program in the Knowing and managing the Brixton first year and the Western Australian Street Wetlands, ed. Bronwen Keighery. Threatened Species and Community Unit, Unpublished report for the Friends of Department of Conservation and Land Brixton Street Wetlands and the Perth Management, provided funding for the Branch of the Wildflower Society. Wild- control program in the second year. The flower Society of Western Australia work at Brixton Street is part of a project (Inc.), Nedlands, Western Australia. managed by the Environmental Weeds Le Maitre, D.C. and Brown, P.J. (1992). Life Action Network and funded through the cycles and fire stimulated flowering in Natural Heritage Trust Bushcare Program geophytes. In ‘Fire in South African and the Lotteries Commission. mountain fynbos’, eds B.W. van Wilgen,