sign in sign up
<<
Home , Acacia dealbata, Acacia mearnsii, Acacia melanoxylon, Acacia pycnantha, Uromycladium tepperianum

Galling guilds associated with dealbata and factors guiding selection of potential biological control agents

R.J. Adair1

Summary The Australian tree Link (Mimosaceae) invades natural ecosystems in both the Northern and Southern Hemispheres, including areas beyond its natural range in . Biological control is under development in South using the -feeding curculionid Melanterius macu- latus Lea. A diverse range of galling insects occur on A. dealbata in Australia, most exhibiting high levels of host specificity and niche partitioning within their host. Galling insects have successfully contributed to biocontrol of other Acacia in . Australian galling insects from A. dealbata have considerable potential for adoption as novel or complementary biocontrol agents. Factors governing the selection of potential agents are considered in the context of impact on the host, efficacy and compatibility with the utilization of the host for timber, pulp, floriculture and fire- harvesting, particularly in resource-poor regions of the world. The potential for biological control of A. dealbata in invaded in Australia is also discussed.

Keywords: wattle, conflict of interest, agent selection.

Introduction of A. dealbata occurs in South Africa where large-scale invasions make other forms of suppression difficult to Silver wattle, Acacia dealbata Link (Mimosaceae: implement. This paper examines the role of classical Botrycephalae), is a widespread and conspicuous tree biological control of A. dealbata using -forming indigenous to forests and woodlands of southeast- agents and how such agents may affect commercial and ern Australia (Costermans, 1983). The species has a utilitarian values of the host tree. broad range with two sub-specific taxa (subsp. dealbata, subsp. subalpina) that are delineated by al- A. dealbata—the invader titude (Kodela and Tindale, 2001). A. dealbata is of- ten abundant in early post-fire vegetation succession, While A. dealbata is native to eastern Australia, ex- where mass germination of soil-stored seed is triggered tensive and expanding naturalized populations occur in by burning. In its native habitat, A. dealbata provides south-west , where the species was ecosystem functions such as food and habitat for fauna introduced for horticultural purposes. Although south- (Broadhurst and Young, 2006) and fixing atmospheric ern Western Australia has an astoundingly rich native nitrogen. The species’ silvery bipinnate foliage and Acacia flora (Hnatiuk and Maslin, 1988), there are no abundant production of bright yellow flowers in winter native Botrycephalae, and very few Western Australian to early spring contributes to the popularity of A. de- are large woody trees. Consequently, invasion albata in horticulture. Large, naturalized populations of A. dealbata into the native vegetation in Western of A. dealbata now occur in many countries and can Australia may have undesirable ecological impacts, al- require management to protect natural and social assets though quantitative impact data both in Australia and (Sheppard et al., 2006; Adair, 2008). Biological control elsewhere are lacking. In South Africa, A. dealbata has been problematic as early as 1915 (Henkel, 1915) and is now a weed of national importance due to negative impacts on water management and conser- 1 Department of Primary Industries, PO Box 48, Frankston, , Australia 3199 . vation (Le Maitre et al., 2002; Nel et al., 2004). More © CAB International 2008 recently, in , A. dealbata was listed as one of the

122 Galling guilds associated with Acacia dealbata and factors guiding selection of potential biological control agents top 20 invasive suggested as targets for biologi- case, biological suppression programs were strongly cal control (Sheppard et al., 2006). Invasions in south- beneficial to the national interest. ern France post-1910 have progressively replaced local Where A. dealbata threatens important assets, bona vegetation including cork , l`arbousier (Arbutus fide utilitarian values need to be taken into account unedo L.) and heather (http://www.worldwidewattle. when designing biological control strategies to reduce com/). A. dealbata is also naturalized in , levels of conflict of interest. Historically, potential con- western , Madagascar, Japan and Chile flicts of interests are avoided by: (1) not initiating bio- (Randall, 2002). logical control programs, (2) undertaking a cost–benefit analysis and proceeding with biological control where Utilitarian values of A. dealbata it is in the public interest, or (3) by targeting specific organs on the host and avoiding negative impacts on In Australia, A. dealbata is utilized in habitat resto- utilitarian interests. ration programs and urban landscaping projects. The species is not utilized commercially, although pollen used by honey bees contributes to the apiary industry. Biological control of Australian acacias In New Zealand and North America, A. dealbata is uti- Classical biological control of Australian acacias lized in horticulture but with limited economic value. was pioneered in South Africa, where eight species In contrast, the exploitation of A. dealbata is well de- are currently subject to active research, development veloped in southern Europe and South Africa where or agent redistribution programs. All of these pro- the species services quite different industries in each grams have succeeded in the establishment of one of these regions. or more agents, and several targets are now subject In Europe, A. dealbata was introduced around 1816 to satisfactory levels of suppression (Dennill et al., (Cavanagh, 2006) where acacias (‘’) are grown 1999; Hoffmann et al., 2002). Two general approach- for horticultural and floricultural purposes. The - ‘mi es to biological control of acacias have been adopted mosa’ cut flower industry in France occupies around in South Africa, each largely governed by the level 200 ha with an estimated value of €3–4 million/year of conflict of interest with commercial or utilitar- (Roland, 2006). Hybrids of A. dealbata and selected ian interests. Economically important species (A. cultivars form the basis of the industry and produce mearnsii, R. Br., A. dealbata, flower crops between December and March. Whether A. decurrens Willd.) are targeted solely for biologi- these selections and hybrids have naturalized in Eu- cal control of reproductive organs with seed-feeding rope is uncertain, but the creation and invasion of de curculionids (Melanterius spp.) that have no nega- novo genotypes by hybridization can complicate clas- tive impacts on vegetative growth of the host . sical biological control programs. Essential oils from In contrast, acacia species of little or no economic the flowers of A. dealbata are used as a fixative and value (Acacia cyclops A. Cunn. ex G. Don., Acacia blending agent in the manufacture of high-grade per- longifolia (Andrews) Willd., (Labill.) fumes and soaps, and the industry consumes around €1 H.L.Wendl., Benth.) are subject to million of refined ‘mimosa’ absolute per year (Roland, biological control of a range of plant organs where 2006). More recently, the French tourism industry has Melanterius spp. or flower-galling Cecidomyiidae promoted the virtues of the ‘Route de Mimosa’ during are used to target reproductive organs along with the main flowering season with numerous festive acti- (Hymenoptera: Pteromlaidae) or Uro- vities linked to this period, undoubtedly contributing to mycladium (Fungi: Uredinales), which gall vegeta- local economies in the Bormes-les- to Grasse tive organs. region. In Australia, biological control of invasive acacia In South Africa, silvicultural operations use Acacia species that have transgressed substantial geographical mearnsii De Wild. And, to a limited extent, Acacia de- barriers (trans-continental invaders) is advocated. A. currens Willd. A. dealbata is not commercially culti- dealbata invasions in Western Australia are suggested vated, but extensive areas of naturalized and invasive as targets for biological control (Adair, 2008). Biologi- populations of A. dealbata in eastern South Africa are cal control of A. longifolia in Portugal has commenced the legacy of early experimental and development pro- following successful control in South Africa (Sheppard grams. Resource-poor communities utilize A. dealbata et al., 2006). for fuel wood, charcoal and construction timber where harvesting is carried out ad hoc and driven by local- ized domestic needs (de Neergaard et al., 2005). The Galling agents and biological control contribution of A. dealbata to the regional and national strategies for Australian acacias economies of South Africa has not been calculated, al- though limited and careful extrapolation from the cost– Galling organisms vary in their impact on the host benefit analysis undertaken for A. mearnsii (de Wit et plant depending largely on the mode of physiological al., 2001) could possibly be made. In the A. mearnsii interaction with the host (innate impact), gall densi-

123 XII International Symposium on Biological Control of Weeds ties, phenological synchronization, location on the host The average dry weight of galled tissue compared and capacity to divert and accumulate resource alloca- to the average weight of the same un-galled organs tion (Hartnet and Abrahamson, 1979; Dorchin et al., was used in this study to indicate the general level or 2006). direction of resource partitioning to gall structures Dennill (1988) outlines ecological hypotheses un- (Table 1). that are heavier than normal un-galled derpinning the successful suppression of A. longifolia organs (positive gall biomass ratio) indicate a possible in South Africa. In this system of ‘forced commitment’, resource sink. Conversely, galled tissues with lower diversion of host resources to gall development occurs biomass than the same organs without galls may indi- at the expense of normal growth functions. In compari- cate the absence of such a resource sink. The were more son, the flower-galling cecidomyiids Dasineura dielsi cecidogenic organisms associated with A. dealbata Rübsaamen and Dasineura rubiformis Kolesik pro- with negative to neutral gall biomass ratios (61%) posed for biological control of Australian acacias induce than those with positive gall biomass ratios (38%; gall structures with biomass and calorific allocations Table 1). the same as or less than normal production. Therefore, disruption to vegetative growth beyond that Selection of potential galling agents created by normal fruit formation is unlikely (Adair, 2005). In such cases where host trees are prevented High costs and safety concerns in the development from producing heavy fruiting loads, vegetative growth and release of biological control agents necessitate was found to be either unaffected or accelerated. This careful selection of organisms destined for detailed process is termed ‘commitment release’ (Adair, 2005) evaluation. Five selection filters are proposed here for and may be applicable to situations where conflicts of pre-screening potential galling agents for A. dealbata, interest are associated with the targeting of vegetative (1) impact efficacy, (2) host specificity, (3) conflict of organs. interest, (4) climatic compatibility and (5) risk of para- In the case of A. dealbata, a diverse assemblage sitism. of galling agents is known with a range of innate im- Impact efficacy: Efficacy filters preceding host speci- pacts. ficity evaluation can effectively narrow the range of organisms for further consideration and may improve Galling biota of A. dealbata the prospects for success (Raghu, et al., 2006). Eco- in Australia logical modelling designed to identify weak points in the host’s life history (Briese, 2006), together with In an extensive survey of Acacia in southern Australia pre-release impact assessment (McClay and Balci- (Adair, 2005), records and accessions of gall-forming unas, 2006), are useful tools for quantitative efficacy insects collected on A. dealbata were extracted and evaluation. However, manipulative techniques to combined with published data records. Thirteen gall- gauge density-based impacts of endophagous organ- inducing species were recorded on A. dealbata: sev- isms, particularly on large woody plants, are some- en restricted to reproductive organs; two restricted to what problematic. While density-related impacts will ; two restricted to stems of various size classes; be important, they remain largely untestable for large one restricted to vegetative and reproductive buds; trees, except perhaps in situations where outbreak and two that attacked a range of host organs (Table 1). populations occur in the agent’s natural range, e.g. More than half of the taxa (61%) belonged to the Ceci- D. rubiformis in Western Australia (Adair, 2005). In domyiidae (Diptera), a family that is well-known from the case of A. dealbata, the innate impacts of galling Australian Mimosaceae (Adair, 2005). organisms form the initial efficacy filter, and organ- All recorded gall-forming biota from A. dealbata isms with impact-class scores of 1 and 2 (Table 1) have restricted host ranges, at least within the Botryceph- are dropped from further consideration. Modelling alae, with the exception of the U. notabile Mc insect reproductive capacity and survivorship predic- Alpine (Uredinales), which is recorded from numerous tions combined with estimation of population densi- bipinnate species of Australian Acacia (Marks et al., ties likely to achieve effective damage to the host may 1982), although host-specific biotypes are known to oc- be the only realistic way of further filtering for efficacy cur within this genus (Morris, 1999). Perilampella sp. of A. dealbata agents. (Pteromalidae) appears to be confined to A. dealbata, Host specificity: Galling organisms are generally host and ?Cecidomyia sp. is restricted to a small group of specific (monophagous) or have a host range restricted closely related Botrycephalae, where A. dealbata is its to closely related plant species (stenophagous); (Anan- principal host. Densities of gall-forming organisms as- thakrishnan, 1984). Nearly all galling agents on A. de- sociated with A. dealbata are generally low, but most albata are stenophagous or polyphagous within Acacia. species are widespread within the natural distribution Host-specificity filtering needs to consider commercial of this species. Dasineura sp. 2 appears to be restricted and utilitarian interests of potentially susceptible non- to south-west . target taxa and should be performed in a regional context

124 Galling guilds associated with Acacia dealbata and factors guiding selection of potential biological control agents - g E E E A A, SA,E A, SA,E A, SA, E A, SA, E A, SA, E for biocontrol Region excluded f − − − − − − + + + + + N N Gall: organ Gall: organ biomass ratio e 1 1 1 1 1 1 1 2 2 2 3 3 3 interest Conflict of d 2 2 2 2 2 1 2 2 2 3 3 3 2 distribution Geographical c 3 3 3 3 3 3 3 2 6 5 5 2 6 class Impact b

A C U, S U, S M,C U, C U, C U, C U, C M, C ?U, S M , ?M, S ?M, C ?M, C Biology a ; 3, distribution uncertain. P P P S S S S S S S S S , polyphagous—found on species within a number of subgenera. M (Ad) A. dealbata Host range Seed Flower bud Flower bud, seed Ovary Ovary Ovary Ovary Pinnule Bud Stem Stem Host organ Leaf rachis Stem, fruit, Australia. in ; 2, widespread across natural range of Pinnule galler Common name Seed galler Pubescent bud Eastern bud- seed galler Lop-sided stem galler Inflated floret galler Fleshy floret galler Hollow galler galler Bud-shoot galler Rachis galler galler galler Galling fungus Small-stem Red plush Acacia dealbata i

A. dealbata sp. sp. sp. 1 sp. 2 sp. 3 h sp. 1 sp. 2 +, gall biomass is greater than host organ. N, gall biomass is approximately equal to host organ; host organ;

by van den Berg (1980, unpublished records) are likely to be the result of misidentification host plant. by van den Berg Africa,; E, Europe. Cecidomyia Undescribed genus Genus species Asphondylia Asphondylia Asphondylia Dasineura pilifera Dasineura Dasineura ? Perilampella ?hecteaeus Perilampella Undetermined notabile Undetermined ; M, monophagous; S, stenophagous—restricted to the Botrycephalae; P A. mearnsii Gall-forming organisms from Gall-forming organisms Acacia dealbata Australia; SA, South −, Gall biomass is lower than U, Univoltine; M, multivoltine; A, alternates between host organs; S, pupation occurs in soil; C, life cycle completed within gall. A, alternates between host organs; U, Univoltine; M, multivoltine; 1, Restricted within natural range of Records from Family Cecidomyiidae Pteromalidae Tetrastichinae Lepidoptera Uredinales A,  2, Minor disruption to normal growth processes but impact unlikely to affect host fitness even in high densities; 3, impact restricted to reproductive organs and gall biomass equal to or less than fruit biomass; 5, moder 5, biomass; fruit than less or to equal biomass gall and organs reproductive to restricted impact 3, densities; high in even fitness host affect to unlikely impact but processes growth normal to disruption Minor 2, ate disruption to normal growth processes; 6, significant processes. interests. commercial most with conflict to high—potential 3. interests; commercial some with conflict to moderate—potential 2, Acacias; Australian with associated interests commercial with conflict not Low—will 1, Ad,

Taxonomic position requires verification using molecular diagnostics, and feeding range assessed no-choice tests. Taxonomic

Table 1. Table a b c d e f g h i

125 XII International Symposium on Biological Control of Weeds by addressing local industry issues. Using natural host Asphondylia sp. 3 that induce single-chambered, thin- records, three regionally based host-specificity filters walled galls, warrant exclusion due to the high prob- are established for Australia, South Africa and Europe. ability of attack by parasitoids. In contrast, the galls In Western Australia, several Botrycephalae acacias are of ?Cecidomyia sp. consist of long compacted hairs cultivated commercially, but none are used by the sil- that surround a hard woody kernel. In Australia, ?Ce- vicultural industries. Based on host specificity, all gall- cidomyia sp. is parasitized by a specialized dipteran ing organisms restricted to the Botrycephalae should (Chloropidae: Gaurax sp.), but remarkably few hy- be considered as potential biocontrol agents for A. de- menopterans, which typically dominate cecidomyiid albata. Only Asphondylia sp. 3 and the polyphagous galls. The parasitism risks of this insect in a biologi- biotypes of Uromycladium should be excluded on this cal control context may be lower than the more simply basis. In South Africa, A. mearnsii and A. decurrens are structured Asphondylia spp. galls. commercially exploited, and galling organisms known from these hosts that have positive gall biomass ratios and are capable of affecting vegetative growth need to Conclusions be excluded. Therefore, Asphondylia sp. 3, a new ceci- The successful suppression of invasive Australian domyiid genus (pinnule galler), Tetrastichinae sp., the acacias using classical biological control has been stem-galling lepidopteran, and the polyphagous bio- achieved through the use of galling agents that induce types of Uromycladium should be excluded in South a debilitating resource allocation commitment in their Africa. In Europe, organisms that attack commercially host (Dennill, 1988) and seed-feeding agents, either in important Botrycephalae acacias are excluded where combination or as a single-agent introduction. A. deal- host structures of importance are affected (Table 1). bata is utilized for commercial and domestic purposes Conflicts of interest: The conflict of interest filter re- in both the Southern and Northern Hemispheres. The lates to direct impacts on the targeted host, A. dealbata. biological control strategy adopted for invasive non- In Europe and South Africa, organisms that directly or commercial acacias in South Africa has limited ap- indirectly (e.g. due to positive gall biomass ratios) af- plication for A. dealbata, except in Western Australia fect structures of importance are excluded. Therefore, where the level of conflicts of interest is low. In other organisms affecting vegetative organs and pre-flower- regions, host- and organ-specific gall-inducing organ- ing structures are excluded as potential agents for Eu- isms known from A. dealbata may contribute to the rope. In South Africa, only organisms with potential to biological suppression of this plant. Control programs affect vegetative growth are excluded. that focus on suppression of seed-producing organs to Climatic compatibility: Close matching be- avoid conflicts of interest need to be guided by the po- tween natural and intended areas of introduction may tential of the agents to achieve ecologically meaningful influence the success of biological control outcomes levels of control. While host impacts induced by en- (Dhileepan et al., 2006). The galling organisms associ- dophagous organisms creating resource sinks on veg- ated with A. dealbata occur over a broad geographic etative growth are difficult to test or predict a priori, area and climate range, including considerable varia- control targets for solely seed-reducing organisms are tion in altitude and rainfall patterns. The only clear more achievable through modelling of the life history exclusion based on climatic considerations was Dasi- attributes and population dynamics of the host. A. deal- neura sp. 2, where known occurrences have a low bata may be a density-independent species, and there- match (using Climex® and Climate®) with introduced fore, suppression by seed-reducing organisms, such as occurrences of A. dealbata in Western Australia. High Melanterius maculatus Lea and Bruchophagus acaciae match levels for this species occur in Europe and South (Cameron) (Hymenoptera) would need to achieve very Africa (unpublished data). high levels of control before population-level impacts Parasitism: Endophagous organisms tend to be suscep- can be obtained. tible to parasitism (Askew, 1980), and failure of some Galling organisms for A. dealbata are available to biocontrol programs using galling agents are attributed contribute to the reduction of reproductive output of to high parasitism levels (Muniappan and McFadyen, A. dealbata, even in situations where the host is com- 2005). Methods for predicting parasitism impacts re- mercially utilized. However, a compatible combina- main elusive (Adair and Neser, 2006). However, gall- tion of agents is more likely to achieve high levels of forming agents that experience low parasitism levels seed reduction than a single agent alone, based on the (<30%) have been successful in suppressing their hosts enormous resource allocation of A. dealbata to flower (Muniappan and McFadyen, 2005). Larger gall size and and fruit production. Seed-feeding organisms that can chamber number can reduce parasitism levels in some find food at low density levels, such as Melanterius gall-forming agents (Manongi and Hoffmann, 1995), ventralis Lea (Donnelly and Hoffmann, 2004), but re- however, this association is not consistent (Waring and spond rapidly to sudden increases in food availability Price, 1989). High endemic parasitism of Australian may work well in combination with organisms that and South African analogues of Asphondylia sp. 2 and attack pre-fruiting stages of the reproductive cycle. The

126 Galling guilds associated with Acacia dealbata and factors guiding selection of potential biological control agents sequence of introduction of combinations of biological Acacia dealbata (Mimosaceae) populations in southeast control agents remains debatable (Impson et al., 2008), Australia. Biological Conservation 133, 512–526. but introductions following a reverse phenological se- Cavanagh, T. (2006) Historical aspects of wattles: the culti- quence (seed-feeders before flower-feeders) may - fa vation of Australian acacias in Great Britain and Europe vour the establishment of organisms that target the end during the 18th and 19th centuries. In: Proceedings Aca- cia2006. Knowing and growing Australian wattles. SGAP of the reproductive process, which could otherwise be Victoria, Australia, pp. 69–82. disadvantaged (Briese, 2006). Costermans, L. (1983) Native trees and shrubs of south-east- A series of five selection filters presented here iden- ern Australia. Rigby, Adelaide, Australia, 422 pp. tifies gall-inducing organisms potentially suitable for Dennill, G.B. (1988) Why a gall former can be a good biocon- suppression of A. dealbata at three levels of conflict trol agent: the gall wasp Trichilogaster acaciaelongifoliae of interest: low (Australia), moderate (South Africa) and the weed Acacia longifolia. Ecological Entomology and high (Europe). Efficacy of impact should precede 13, 1–9. other selection filters (Raghu et al., 2006), and while Dennill, G.B., Donnelly, D., Stewart, K. and Impson, F. difficult to quantify for organisms restricted to repro- (1999) Insects used for the biological control of Austra- ductive organs on large perennial trees, manipulative lian Acacia species and lophantha (Willd.) techniques are technically possible (Balciunas and Nielsen () in South Africa. African Entomologi- cal Memoir 1, 45–54. Burrows, 1993). de Neergaard, A., Saarnak, C., Hill, T., Khanyile, M., Berosa, A.M., -Thomsen, T. (2005) Australian wattle spe- Acknowledgements cies in the Drakensberg region of South Africa—an in- vasive alien or a natural resource? Agricultural Systems I thank S. Neser, N. Dorchin and S. Raghu for valu- 85, 216–233. able contributions that assisted in improving an ear- de Wit, M.P., Crookes, D.J. and van Wilgen, B.W. (2001) lier draft of this paper. John Weiss performed climate Conflicts of interest in environmental management: esti- matches for Dasineura sp. 2. The Department of Water mating the costs and benefits of a tree invasion.Biological Affairs and Forestry, South Africa (Working for Water Invasions 3, 167–178. Dhileepan, K., Wilmont Senaratne, K.A.D. and Raghu, S. Program) contributed to the survey of gall-forming or- (2006) A systematic approach to biological control agent ganisms on Australian acacias. David Yeates (ANIC) exploration and prioritisation for prickly acacia (Acacia kindly identified the Chloropid parasitoid. nilotica ssp. indica). Australian Journal of Entomology 45, 303–307. Donnelly, D. and Hoffmann, J.H. (2004) Utilisation of an un- References predictable food source by Melanterius ventralis (Cole- optera: Curculionidae), a seed-feeding biological control Adair, R.J. (2005) Seed-reducing Cecidomyiidae as potential agent of Acacia longifolia (Mimosaceae) in South Africa. biological control agents for invasive Australian wattles in Biocontrol 49, 225–235. South Africa, particularly and A. cyclops. Dorchin, N., Cramer, M. and Hoffmann, J.H. (2006) Photo- PhD Thesis, University of Cape Town, South Africa. synthesis and sink activity of wasp-induced galls in Aca- Adair, R.J. (2008) Biological control of Australian native cia pycnantha. Ecology 87, 1781–1791. plants, in Australia, with an emphasis on acacias. Muel- Hartnet, D.C. and Abrahamson, W.G. (1979) The effect of leria 26, 67–78. stem gall insects on life history patterns in Solidago ca- Adair, R.J. and Neser, O. (2006) Gall-forming Cecidomyiidae nadensis. Ecology 60, 910–917. from acacias: can new parasitoid assemblages be predict- Henkel, J.S. (1915) Report of the Conservator of Forests. ed? In: Ozaki, K., Yukawa, J., Ohgushi, T. and Price, P.W. Department of Water Affairs and Forestry, South Africa, (eds) Galling Arthropods and their Associated Ecology pp. 16–17. and evolution. Springer-Verlag, Tokyo, pp. 133–147. Hnatiuk, R.J. and Maslin, B.R. (1988) Phytogeography of Ananthakrishnan, T.N. (1984) Adaptive strategies in ceci- Acacia in Australia in relation to climate and species rich- dogenous insects. In: Ananthakrishnan, T.N. (ed.) Biology ness. Australian Journal of Botany 36, 361–383. of Gall Insects. Oxford and IBH Publishing, New Delhi, Hoffmann, J.H., Impson, F.A.C., Moran, V.C. and Donnelly, India, pp. 1–9. D. (2002) Trichilogaster gall wasps (Pteromalidae) and Askew, R.R. (1980) The diversity of insect communities in biological control of invasive golden wattle trees (Acacia leaf mines and plant galls. Journal of Animal Ecology 49, pycnantha) in South Africa. Biological Control 25, 64–73. 817–829. Impson, F., Moran, V.C., Kleinjan, C., Hoffmann, J.H. and Balciunas, J.K. and Burrows, D.W. (1993) The rapid suppres- Moore, J.A. (2008) Multiple species introductions of sion of the growth of Melaleuca quinquenervia saplings biological control agents against weeds: look before you in Australia by insects. Journal of Aquatic Plant Manage- leap. In: Proceedings of the 12th International Symposium ment 31, 265–270. on Biological Control of Weeds (this volume). Briese, D.T. (2006) Can an a priori strategy be developed for Kodela, P.G. and Tindale, M.D. (2001) Acacia dealbata biological control? The case of Onopordum spp. thistles in subsp. subalpina (Fabaceae: ), a new subs- Australia. Australian Journal of Entomology 45, 317–323. pecies from south-eastern Australia. Telopea 9, 319–322. Broadhurst, L.M. and Young, A.G. (2006) Reproductive Le Maitre, D.C., van Wilgen, B.W., Gelderblom, C.M., Bai- constraints for the long-term persistence of fragmented ley, C., Chapman, R.A. and Nel, J.A. (2002) Invasive alien

127 XII International Symposium on Biological Control of Weeds

trees and water resources in South Africa: case studies of Nel, J.A., Richardson, D.M., Rouget, M., Mgidi, T.N., the cost and benefits of management. Forest Ecology and Mdzeke, N., Le Maitre, D., van Wilgen, B.W., Scho- Management 160, 143–159. negevel, L., Henderson, L. and Neser, S. (2004) A pro- Manongi, F.S. and Hoffmann, J.H. (1995) The incidence of posed classification of invasive plant species in South parasitism in Trichilogaster acaciaelongifoliae, a gall- Africa: towards prioritising species and areas for man- forming biological control agent of Acacia longifolia in agement action. South African Journal of Science 100, South Africa. African Entomology 3, 147–151. 53–64. Marks, G.C., Futher, B.A. and Walters, N.E.M. (1982) Tree Raghu, S., Wilson, J.R. and Dhileepan, K. (2006) Refining diseases in Victoria. Forests Commission Victoria, Mel- the process of agent selection through understanding plant bourne, Australia, 149 pp. demography and plant response to herbivory. Australian McClay, A. and Balciunas, J.K. (2006) The role of pre-release Journal of Entomology 45, 308–316. efficacy assessment in selecting classical biocontrol agents Randall, R. (2002) A Global Compendium of Weeds. R.G. and for weeds—applying the Anna Karenina principle. Bio- F.J. Richardson, Melbourne, Australia, 905 pp. logical Control 35, 197–207. Roland, W.A. (2006) Australian acacias in Europe. In: Pro- Morris, M.J. (1999) The contribution of the gall-forming rust ceedings Acacia2006. Knowing and growing Australian fungus (Sacc.) McAlp. to the wattles. SGAP Victoria, Australia, pp. 83–86. biological control of Acacia saligna (Labill.) Wendl. (Fa- Sheppard, A.W., Shaw, R.H. and Sforza, R. (2006) Top baceae) in South Africa. African Entomology Memoir 1, environmental weeds for classical biological control in 125–128. Europe: a review of opportunities, regulations and other Muniappan, R. and McFadyen, R.E. (2005) Gall-inducing barriers to adoption. Weed Research 46, 93–117. arthropods used in the biological control of weeds In: Ra- Waring, G.L. and Price, P.W. (1989) Parasitoid pressure and man, A. Schaefer, C.W. and Withers, T.M. (eds) Biology, the radiation of a gall forming group (Cecidomyiidae: ecology, and evolution of gall-inducing arthropods, Vol 2. Asphondylia spp.) on creosote bush (Larrea tridentata). Science Publishers, Enfield, NH, pp. 709–730. Oecologia 79, 293–299.

128