Plant Protection Quarterly Vol.26(2) 2011 47 populations where they are invading na- tive plant communities. Specifi cally, we Seed dynamics of the invasive geophyte Lachenalia investigated seed viability and persistence in the soil seed bank of L. refl exa. refl exa Thunb. in south-west Australia Methods Kate Brown and Grazyna Paczkowska, Department of Environment and Study area and seed burial site Conservation, Swan Region, PO Box 1167, Bentley Delivery Centre, Western The site chosen for the burial of L. refl - Australia 6983, Australia. exa seed was on the Swan Coastal Plain in Woodvale Nature Reserve (31°47'03"S, 115°46'55"E), within close proximity to the city of Perth, Western Australia. The dis- turbed Banksia attenuata, Eucalyptus wood- land on the Spearwood Dune System of the reserve are species rich, (average of Summary and Longman 2004). Protection and res- 55.2 species in a 10 × 10 m quadrat) with Geophytes from South Africa’s Cape toration of remnant vegetation across the an understorey of shrubs, herbs, sedges Province are a particularly serious group region often requires an understanding of and grasses (Gibson et al. 1994). Dense of invasive plants in south-west Aus- how to manage invasive geophytes. populations of L. refl exa were present in tralia. They threaten natural biodiversity Lachenalia, in the family Asparagaceae, the understorey of the woodlands at the across the region, invading relatively is a genus of around 120 species of winter seed burial site. undisturbed habitat and displacing na- growing geophytes endemic to South Af- tive plant communities. Their success as rica and Namibia (Duncan et al. 2005). Na- Initial seed viability invaders has been linked to rapid and tive to the Cape Floristic Region, Lachenalia Seed was collected at Woodvale Nature profuse seedling germination, however refl exa (yellow soldier) was fi rst recorded Reserve from mature fruits of approxi- studies on a number of species have doc- as naturalized in south-west Australia mately 200 plants scattered through natu- umented short lived soil seed banks. This in 1938 (Western Australian Herbarium ralized populations of L. refl exa in Novem- study investigates seed viability and soil 1998–2010). It is now a serious weed of ber 2004. seed bank persistence of the Cape geo- Banksia/Eucalyptus woodlands on the Initial seed viability was determined phyte, Lachenalia refl exa Thunb. where western side of the Swan Coastal Plain in germination tests carried out in Janu- it is invading Banksia woodlands on the and has recently been the target of a re- ary 2005. Five replicates of 25 seed were Swan Coastal Plain, Western Australia. gional weed strategy in south-west Aus- placed on agar (0.75%) and placed in Seed was collected from naturalized tralia (Bettink 2009). With a single annu- growth cabinets at 15°C with a 12 hour populations in the summer of 2004/2005 ally renewed bulb, L. refl exa disperses into photo period. Seeds that did not germi- and placed in nylon mesh bags buried at bushland mainly by small (2 mm) shiny nate after 10 weeks were assessed for vi- two different soil depths, in replicated black seed that have a tiny arilode. Papery ability using a cut test. Seeds were scored plots across fi eld sites (February 2005). capsules contain 60–100 seed and there as viable if a fi rm, white, moist endosperm Samples were retrieved and assessed in are up to 10 capsules per plant. Seed is was present (Ooi et al. 2004). May, July and September 2005, Septem- released in late spring or summer and ap- ber 2006 and September 2007. pears to disperse via water into bushland Soil seed bank persistence Initial tests found 95% seed viability. following autumn rain. Seed was placed in 15 × 20 cm nylon mesh More than 90% germinated or rotted in Although appropriate control methods bags containing 200 g sterile sand to keep the fi rst autumn/winter. No viable seed are well documented (Brown et al. 2002, individual seed separated, and buried in remained at the 10 cm depth by Septem- Brown and Brooks 2003) there is little February 2005 in fi ve plots placed across ber 2006 or at the 1 cm depth by Septem- known about ecology or reproductive bi- naturalized populations of L. refl exa. In ber 2007. The results show that L. refl exa ology where L. refl exa is invading native each plot 20 bags were buried, each bag seed have high initial viability but do plant communities. In particular, seed via- containing 500 seeds. Ten bags were bur- not persist in the soil seed bank for more bility, patterns of seedling emergence and ied in the top 1 cm of soil and ten bags than three years. persistence of naturalized populations in were buried at 10 cm. These depths were Keywords: Invasive, geophyte, germi- the soil seed bank are not well understood. chosen to refl ect natural seed burial depths nation, viability, soil seed bank, Lache- This information is critical for understand- following dispersal with 1 cm represent- nalia refl exa. ing how to manage invading populations ing superfi cial surface burial versus burial over time including the length of any con- resulting from substantial soil disturbance Introduction trol program. (10 cm). Chicken wire was secured over A detailed understanding of the biology of There are few detailed studies on the each plot to protect the site from rabbit invasive plant species underpins effective seed ecology of individual species of and bandicoot diggings. The soils across weed management strategies (Hobbs and geophytes where they occur naturally in the site were uniform grey to orange light Humphries 1995, Mortensen et al. 2000). South Africa. However it is generally as- sand. Knowledge of seed biology in particular sumed that seed of many geophytes are One bag per soil depth was recovered provides an understanding of patterns in not persistent in the soil seed bank (Man- from each plot in May, July and September recruitment of an invasive species and the ning et al. 2002, Keeley et al. 2006). There is of 2005, September 2006 and September potential persistence of populations over also evidence that for some species seed- 2007. Seeds recovered in May and July time. ling recruitment and population expan- were scored as either potentially viable With the capacity to invade undisturbed sion occurs in a small window following if they were fi rm and intact or as germi- bushland and form dense monocultures, fi re (Le Maitre and Brown 1992, Keeley nated/rotted if they were not. It was not South African geophytes are a particularly and Bond 1997). possible to separate germinated from rot- serious group of invasive plants in south- The objective of this study was to in- ted seed. Following retrieval in Septem- west Australia. A total of 53 species have vestigate the seed biology of L. refl exa and ber 2005, 2006 and 2007, seed that were been recorded as naturalized (Keighery gain an understanding how to manage still intact were assessed for viability in 48 Plant Protection Quarterly Vol.26(2) 2011 growth cabinets at 15°C with a 12 hour 600 photo period. Seed that did not germinate after 10 weeks were assessed using a cut test and were scored as viable if a fi rm, 500 white, moist endosperm was present. 400 Results Initial germination tests indicated seed 1 cm was 95.2% viable with all germination oc- 300 curring within 18 days. 10 cm Of the seed buried at 1 cm, 18.8% (±1.3) were still viable when retrieved in May 200 2005 and 8.74% (±3.8) were still viable in July. By September 2005, 3.3% (±1.5) re- 100 mained viable. In September 2006 1.3% seed viable of Number (±1.1) remained viable and by September 2007 no seed of L. refl exa buried at 1 cm 0 remained viable. Feb-05 May-05 Jul-05 Sep-05 Sep-06 Sep-07 Of the seed buried in the fi eld in Febru- ary 2005 at 10 cm, 5.96% (±1.9) were still viable when retrieved in May and 0.24% Figure 1. Number of seed that remained viable buried at 1 cm and 10 cm (±0.2) were still viable in July. By Septem- over three years. ber 2005, 0.04% (±0.4) was still viable. By September 2006 bags buried at 10 cm con- tained no viable seed (Figure 1). seed bank persistence of native geophytes California chaparral. International Jour- Discussion of the Californian chaparral (Keeley 1994, nal of Wildland Fire 18 (5), 623-30. The results of our study indicate that the Tyler and Borchert 2002, Keeley et al. Brown, K. and Brooks, K. (2003). ‘Bush- seed from naturalized populations of L. 2006). land weeds: a practical guide to their refl exa has high viability (95.2%) with ger- While seed plays a crucial role in the management’. (Environmental Weeds mination occurring relatively rapidly, all recruitment of individuals and expansion Action Network, Greenwood). within 18 days, under controlled condi- of populations, the persistence of popula- Brown, K. and Paczkowska, G. (2008). tions. With more than 80% of seed ger- tions over time depends on the annually Seed biology of two invasive South Af- minating/rotting within three months of renewed bulbs of L. refl exa. While care- rican geophytes and implications for burial (May 2005) in the fi eld, indications fully targeted, selective herbicide appli- natural area management. Ecological are that germination also occurs relatively cation over one or two years will result Management and Restoration 9 (3), 232- rapidly under fi eld conditions. The pat- in effective control of bulbs (Brown et al. 34. tern of high viability and rapid germina- 2002, Brown and Brooks 2003), the results Brown, K., Brooks, K., Madden, S. and tion is similar to two other invasive geo- of this research shows managers will need Marshall, J. (2002). Control of the exotic phytes in south-west Australia, Freesia alba to survey and control germinating seed- bulb, yellow soldier (Lachenalia refl exa) × leichtlinii (Iridaceae) and Sparaxis bulb- lings for at least three years. invading a Banksia woodland, Perth, ifera (Iridaceae) (Brown and Paczkowska A non-persistent seed bank implies Western Australia. Ecological Manage- 2008). These results also support those limited opportunities for seedling recruit- ment and Restoration 3 (1), 26-34. of van Kleunen and Johnson (2007) that ment. Where they occur in Mediterranean Duncan, G.D., Edwards, T.J. and Mitch- South African geophytes naturalized out- ecosystems geophytes often fl ower par- ell, A. (2005). Character variation and a side their native range tend to have rapid ticularly prolifi cally following fi re. This cladistic analysis of the genus Lachenalia and profuse seedling germination. is often followed by seedling recruitment Jacq.f. ex Murray (Hyacinthaceae). Acta After seven months of burial, at the in the second year following fi re (Kruger Horticulturae 673 (1), 113-20. end of the fi rst growing season (spring 1983, Le Maitre and Brown 1992, Keeley Gibson, N., Keighery, B.J., Keighery, G.J., 2005) almost all the seed had germinated et al. 2006, Borchert and Tyler 2009). The Burbidge, A.H. and Lyons, M.N. (1994). or rotted. After 14 months (spring 2006) prolifi c amount of seed produced is able to A fl oristic survey of the southern Swan no viable seed of L. refl exa remained at 10 germinate and establish particularly well Coastal Plain. Report for the Austral- cm and after 21 months (spring 2007) no in the post fi re environment. The second ian Heritage Commission prepared by viable seed remained at 1 cm. While L. re- growing season following a summer fi re, the Department of Conservation and fl exa does not form a persistent soil seed when seed produced from post fi re fl ow- Land Management and the Conser- bank, its seed persists longer in the soil ering would germinate, could be a crucial vation Council of Western Australia than two other South African geophytes time to control L. refl exa and prevent ex- (Inc). that are invasive in south-west Western pansion of existing populations. Hobbs, R.J. and Humphries, S.E. (1995). Australia, Freesia alba × leichtlinii and An integrated approach to the ecology Sparaxis bulbifera. The seed of both these References and management of plant invasions. species was found to persist for just one Bettink, K. (2009). Yellow soldier (Lachena- Conservation Biology 9 (4), 761-70. season (Brown and Paczkowska 2008). lia refl exa). Draft strategic control plan Keeley, J.E. (1994). Seed-germination pat- The lack of a persistent soil seed bank in for the Swan NRM. http://www.dec. terns in fi re-prone Mediterranean cli- these three invasive geophytes supports wa.gov.au/content/view/3582/2024/ mate regions. In ‘Ecology and bioge- evidence from their native range of Cape (accessed March 2010). ography of Mediterranean ecosystems Floristic Region, that geophytes typically Borchert, M. and Tyler, C.M. (2009). Pat- in Chile, California and Australia’, eds possess non-dormant seeds (Keeley and terns of post-fi re fl owering and fruit- M.T.K. Arroyo, P.H. Zedler and M.D. Bond 1997, Manning et al. 2002). It is also ing in Chlorogalum pomeridianum var. Fox, pp. 239-273. (Springer-Verlag, consistent with fi ndings of studies on soil pomeridianum (DC.) Kunth in southern New York). Plant Protection Quarterly Vol.26(2) 2011 49 Keeley, J.E. and Bond, W.J. (1997). Conver- gent seed germination in South African fynbos and Californian chaparral. Plant Economic and environmental assessment of the Ecology 133, 153-67. Keeley, J.E., Fotheringham, C.J. and performance of reduced rates of two post-emergence Baer-Keeley, M. (2006). Demographic patterns of post-fire regeneration in herbicides in an arid irrigated production system of Mediterranean-climate shrublands of central Australia: a pilot study California. Ecological Monographs 76, 235-55. Martin Hidalgo, Glen Oliver and S. Raghu, Arid Zone Research Institute, Keighery, G.J. and Longman, V. (2004). Department of Resources, Northern Territory Government, PO Box 8760, The naturalized vascular plants of Alice Springs, Northern Territory 0871, Australia. Western Australia: checklist, environ- mental weeds and distribution in IBRA regions. Plant Protection Quarterly 19, 12-32. Kruger, F.J. (1983). Plant community di- versity and dynamics in relation to fi re. Summary In ‘Mediterranean-type ecosystems: The use of herbicides can add consid- to the economy (ABS 2009a,b). The cost the role of nutrients’, eds F.J. Kruger, erable costs to production practices on of non-fertilizer chemical inputs (e.g. pes- D. Mitchell and J.J. Jarvis, pp. 446-72. marginal land, and increase risks to en- ticides, soil amendments) totalled nearly (Springer-Vorlage, Berlin). vironmental and human health. The rela- $3 billion for the period 2006–2007 (ABS Le Maitre, D.C. and Brown, P.J. (1992). tive performance of label and sub-label 2008). While such agricultural activity has Life cycles and fi re stimulated fl ower- rates of the post-emergence herbicides successfully increased yields of food crops ing in geophytes. In ‘Fire in South Af- (Amitrole T and Basta) and label rates of this productivity has come at an environ- rican mountain fynbos’, eds B.W. van Roundup was compared within a ben- mental (e.g. impact on ecosystem services, Wilgen, D.M. Richardson, F.J. Kruger efi t-cost analysis framework, in an arid pollution) and social cost (e.g. health risks and H.J. van Hensbergen, pp. 145-60. irrigated production system in central to rural communities from chemical use), (Springer-Verlag, Berlin). Australia. Sub-label rates of Amitrole raising concerns about continuing such an Manning, J., Goldblatt, P. and Snijman, T and Basta were as effective at weed approach unchecked into the future (Bel- D. (2002). ‘The colour encyclopaedia of suppression as their label rates (LR); ap- lamy and Johnson 2000, Matson et al. 1997, Cape bulbs’. (Timber Press, Oregon). proximately 50% reduction in weed cover Tilman et al. 2002). Mortensen, D.A., Bastiaans, L. and Sat- was recorded across the trial with both Reducing the reliance on non-fertiliz- tin, M. (2000). The role of ecology in label and sub-label rates for both her- er chemical inputs by agriculture is not the development of weed management bicides. The sub-label rates of Amitrole only environmentally benefi cial, but is systems: an outlook. Weed Research 40 T (75% LR and 65% LR) and Basta (75% also economically prudent. For example, (1), 49-63. LR) had a statistically similar economic the cost of herbicide use for 2006–2007 in Ooi, M., Auld, T. and Whelan, R. (2004). benefi t-cost ratio as their corresponding Australian agriculture was nearly $980 Comparison of the cut and tetrazo- LRs, and the LR of Roundup. The equiv- million (ABS 2008). However, numerous lium tests for assessing seed viability: a alence of Basta’s sub-label rate in terms studies have shown that the optimal rates study using Australian native Leucopo- of economic effi ciency is even more note- of application of herbicides vary in rela- gon species. Ecological Management and worthy if one takes into account that sig- tion to environmental context (Bostrom Restoration 5 (2), 141-3. nifi cantly lower amounts of the herbicide and Fogelfors 2002) and that in many Tyler, C. and Borchert, M. (2002). Repro- needed to be applied to achieve the level instances the required application rate duction and growth of the chaparral of weed suppression obtained using the of such chemical interventions are below geophyte, Zigadenus fremontii (Liliace- label rate of Roundup. Our results sug- that recommended by chemical manufac- ae), in relation to fi re. Plant Ecology 165, gest ways to improve the economic and turers (Zhang et al. 2000). Identifi cation 11-20. environmental effi ciencies of herbicide of the level of chemical intervention for van Kleunen, M. and Johnson, S.D. (2007). use in the arid, irrigated production sys- a given production context can therefore South African Iridaceae with rapid and tems of central Australia. result in chemical use that is both effective profuse seedling emergence are more Keywords: Sub-label rates, buffel at achieving the economic objective, and likely to become naturalized in other grass, arid-zone weeds, cost-benefi t anal- reducing risks to the environment. regions. Journal of Ecology 95, 674-81. ysis, sustainability. Primary production in central Aus- Western Australian Herbarium (1998– tralia is limited by aridity, nutrient-poor 2010). FloraBase—the Western Austral- Introduction soils, and remoteness that restricts ac- ian fl ora. Department of Environment The global trend in farming practices cess to markets and labour; however, the and Conservation. http://florabase. since the 1950s is one that is increasingly availability of aquifer resources (albeit of dec.wa.gov.au/ (accessed May 2010). reliant on intensive and extensive chem- variable quality) makes sustainable irri- ical-inputs in the form of fertilizers, and gated agriculture feasible (Slatyer 1961). pesticides (Matson et al. 1997, Tilman et While the inherent productivity of soils al. 2002). Productivity in Australian agri- may make chemical control of weeds eco- culture mirrors global trends, increasing nomically viable in high quality agricul- by nearly 250% over the past half a cen- tural lands, the use of herbicides is often tury, a fact at least partly attributable to a signifi cant production cost in marginal, reliance on chemically intensive farming irrigation-reliant, arid-zone production (Llewellyn et al. 2002, Radcliffe 2002). Agri- systems of central Australia, and is typi- culture currently occupies approximately cally only viable in high-value crops (Ellis 54% of the total Australian land area, con- et al. 2010, Faroda et al. 2007, Heong et al. tributing nearly $43.27 billion per annum 1995). As in other arid locations, weeds