Biological Control of Melaleuca Quinquenervia: Goal-Based Assessment of Success

Biological control of Melaleuca quinquenervia: goal-based assessment of success

T.D. Center,1 P.D. Pratt,1 P.W. Tipping,1 M.B. Rayamajhi,1 S.A. Wineriter2 and M.F. Purcell3

Summary Success means different things to different people. Unfortunately, the success or failure of weed biological control projects is often evaluated by nonparticipants lacking knowledge of the original goals set by project architects. Criteria for success should match objectives and goals clearly articulated so that success can be properly archived for future synthesis. The Australian tree Melaleuca quinquenervia (Cav.) S.T. Blake, an aggressive invader of the Florida Everglades, may be the largest plant ever targeted for biological control. We realized early on that biological control agents would not remove the many tons of woody biomass that comprised these infestations and so would be unlikely to reduce the infested acreage. Control of this plant by other means, however, was complicated by the billions of canopy-held seeds that are released following injury to the tree. A plan was developed in coordination with land management agencies wherein the goal of biological control was to curtail melaleuca expansion and suppress regeneration while using other means to remove mature trees. Three insect species have been released and others are under consideration. These agents, supplemented by the impacts of an adventive rust fungus and a scale insect, have met established goals and this project shows signs of an emerging success based on the established goals.

Keywords: Everglades, invasive plants, habitat restoration.

Introduction stead, biological control is often the method of last re- sort after other methods against recalcitrant weeds have ‘Success has many fathers while failure dies an or- failed. This is not conducive to improving the overall phan’. This oft-quoted aphorism illustrates the political statistical success rate but is often the most responsible necessity of highlighting successes when they occur so or economic option. Biological control of many seri- that one’s endeavors continue to be supported in the ous weed problems would likely never be attempted if future. Unfortunately, weed biological control proj- target choice was based primarily on maximizing the ects are rarely undertaken based on the likelihood of a probability of success as advocated by Peschken and successful outcome (Peschken and McClay, 1995). In- McClay (1995). Many investigators have focused on the performance of individual agents to gauge success, primarily 1 US Department of Agriculture, Agricultural Research Service, Invasive with an aim toward predicting which taxonomic groups Plant Research Laboratory, 3225 College Avenue, Fort Lauderdale, FL 33312, USA. make the best biological control agents. Such post-hoc 2 US Department of Agriculture, Agricultural Research Service, Inva- analyses suggest a low success rate for weed biologi- sive Plant Research Laboratory, 1911 SW 34th Street, Gainesville, cal control, with only a small proportion of success- FL 32608, USA. fully established agents producing effective control 3 US Department of Agriculture, Agricultural Research Service, Austral- ian Biological Control Laboratory, and Commonwealth Scientific and (Crawley, 1989a,b). Critics have used these statistics to Industrial Research Organization, Entomology, Long Pocket Labora- advise against the use of biological control as a weed tories, 120 Meiers Road, Indooroopilly, QLD 4068, Australia. management tool (Louda and Stiling, 2004). McFa- Corresponding author: T.D. Center . this advice and emphasized the need for project-based © CAB International 2008 assessments.

655 XII International Symposium on Biological Control of Weeds

Most authors use the term ‘success’ to refer only Control of melaleuca is complicated by the fact that to ‘complete success’, wherein no other measures are it grows in areas that are hazardous and strenuous to needed to reduce the weed populations to acceptable lev- work in and where access is difficult. These difficul- els. However, this neglects the importance of partially ties are exacerbated by the tree’s reproductive biology. successful projects that have value when less effort Melaleuca flowers numerous times each year, often is subsequently required to control the weed, because several times on the same stem axis due to indetermi- the density or extent of weed populations is reduced, nant growth, forming spike-like clusters composed of or the weed is less able to reinvade cleared areas or multiple, dichasial groups of three flowers each (Tom- is slower to disperse (Hoffmann, 1995). Success and linson, 1980). Each cluster contains up to 75 individual failure are at the extreme ends of a continuum of many flowers. Fruits arising from these flowers are persistent possible outcomes and even moderate amounts of serotinous capsules that each contains 200–350 minute stress can reduce the competitive ability of a weed and seeds (Meskimen, 1962). These generally remain in the render it less invasive (Center et al., 2005; Coetzee et fruits until disruption of the vascular connection causes al., 2005). the capsules to desiccate and open, often en masse, Successful biological control agents often act by after a fire, freeze, drought, or herbicide treatment preventing continued expansion of a weed population, (Meskimen, 1962). A few (about 12% per year) open rather than by reducing population densities (Hoffmann, continuously, as radial growth of the stem separates the these proceedings). Hoffmann also noted that it may be vascular connection, producing a light, perpetual seed necessary to model weed outbreaks that never happen rain of about 3 billion seeds/ha/year (M. Rayamajhi to perceive biological control effects. Documentation et al., unpublished data). Seeds that fall to the ground of such effects is difficult, at best, which explains why form a rather short-lived soil seed bank with a half-life so many projects are incompletely evaluated and even of less than 1 year (Van et al., 2005). A single large successful projects may be undervalued or forgotten. tree located within a dense stand retains about 50 mil- Thus, statistical success rates should be viewed with lion seeds in its canopy with stands holding as many circumspection, inasmuch as only obvious successes as 25 billion seeds/ha (M. Rayamajhi, unpublished are reported. Furthermore, weed declines may occur in- data). An isolated tree may hold twice as many seeds crementally over many years or even decades and may as one of similar size in a dense stand. Surprisingly, a not be easily observed, especially when observational large proportion (85–90%) of these are actually hollow baselines shift over time, project funding terminates, or seed coats (Rayachhetry et al., 1998; Rayamajhi et al., personnel changes interrupt collection of critical data. 2002). Nonetheless, the remaining 10–15% of embry- Success of projects should be assessed in terms of onic seeds create an enormous regenerative capacity the project’s original goals and objectives. Hence, a capable of producing seedling densities of up to 2256 project can and should be deemed successful whether seedlings/m2 (Franks et al., 2006) following a massive or not the density of the weed is reduced so long as the simultaneous seed shed induced by fire, drought, or goals set out by the project architects are met. In this herbicide application. These may grow into thickets of sense, it is possible to have complete success without up to 130,000 small trees/ha (Van et al., 2000). As the complete control so long as the project goals are clearly stand matures, it thins to about 8000–15,000 trees/ha stated, understood and documented. A recent proj- comprised mostly of mature trees with an understory ect aimed at the control of Melaleuca quinquenervia of suppressed saplings (Rayachhetry et al., 2001). The (Cav.) S.T. Blake (melaleuca) in South Florida as part standing biomass in these forests has been estimated at of a broader Everglades restoration effort is used herein 129–263 metric tonnes (t)/ha (Van et al., 2000). to illustrate this concept. Isolated individual trees constitute a seed source for further encroachment. The seeds, when released, gener- The target ally fall within 15 m of the parent tree (Meskimen, 1962). They often grow into dome-shaped clumps or ‘heads’ Melaleuca is a large tree (up to 30 m tall) of Australian with the parent trees in the centre and progressively origin that was introduced into southern Florida during younger trees toward the periphery. These eventually the latter part of the 19th century (Dray et al., 2006). It coalesce with others blanketing vast acreages of wet- has invaded wetland habitats, especially fire-maintained lands with dense swamp forests. The isolated ‘outliers’ Everglades ecosystems, where the burning regime now therefore are regarded as potential new infesta- favours melaleuca over less fire-tolerant native species. tions and, as part of a quarantine strategy, are first As a result, vast areas of these heterogeneous marshes priority for control operations (Woodall, 1981). have been transformed into swamp forests consisting of The trees within these stands produce multiple ad- melaleuca monocultures. Melaleuca rapidly dominates ventitious roots that form an intertwined skirt at the infested areas after its initial colonization (Laroche and waterline or on saturated soil (Meskimen, 1962). These Ferriter, 1992) and at its peak was estimated to occupy contribute biomass to the forest floor and trap large at least 607,000 ha of conservation lands in the south- amounts of litterfall as well as organic debris causing ern part of Florida (Bodle et al., 1994). soil accretion (White, 1994), thus increasing the local

656 Biological control of Melaleuca quinquenervia: goal-based assessment of success elevation (T. Center, personal observation). Altering trol would take time, whereas chemical and mechanical the elevation of the Everglades even by a few centime- control could be employed rapidly. So the plan relied ters dramatically shifts plant community composition on an early deployment of traditional control measures (Ogden, 2005), thus these newly created melaleuca is- that would gradually be supplanted by biological con- lands forever change the physiography and ecology of trol as agents became available (Figure 1). the area. There is also evidence that essential oils in melaleuca litter may be allelopathic (Di Stefano and The biological control agents Fisher, 1983). These changes render infested habitats unsuitable for many native species making restoration Insects associated with melaleuca were enumerated in difficult if not impossible. Australia during the late 1980s and early 1990s (Bal- ciunas, 1990). These inventories revealed an entomo- The melaleuca management plan fauna of over 400 species (Balciunas, 1990; Balciunas et al., 1993a,b, 1994, 1995a,b,c; Burrows et al., 1994, The South Florida Water Management District in con- 1996). The most promising species were studied fur- junction with the Exotic Pest Plant Council convened ther and three have now been released. a meeting of the major agencies that were managing The first insect evaluated was the weevilOxyops vi- the melaleuca problem. They developed a ‘Melaleuca tiosa Pascoe (Purcell and Balciunas, 1994). This insect, Management Plan for Florida’, published during 1990 being a flush feeder on growing stem tips, was desir- and revised in 1994 and 1999. able because of its ability to disrupt flower production, Two points were evident during the development of which depends on continual growth of the stem axis. this plan. First, biological control could not eliminate It proved to be host-specific (Balciunas et al., 1994; the huge amounts of woody biomass present; herbicidal G. Buckingham, unpublished report) and was released and mechanical control would therefore be needed to during 1997 (Center et al., 2000). Its need to pupate reduce the infestations to a maintenance level. Second, in dry soil (Purcell and Balciunas, 1994; Center et al., public agencies could not expend public funds to con- 2000), however, limited it to habitats that were not per- trol melaleuca infestations on private lands that often manently under water. Field and laboratory assessments abutted cleared tracts of public lands. These unassail- of a mirid, Eucerocoris suspectus Distant, in Austra- able stands provided an invasion front and a potential lia (Burrows and Balciunas, 1999) suggested that its seed reservoir to support reinvasion of cleared sites. The host range was limited to melaleuca and a few close role of biological control in this plan was to neutralize relatives. Follow-up studies in US quarantine facilities the reproductive potential of these remaining stands by failed to confirm this so it was dropped from consider- reducing seed production, seedling recruitment and re- ation. The host range of the pergid sawflyLophyrotoma generation; thereby inhibiting spread, reducing reinva- zonalis (Rohwer) proved sufficiently narrow (Burrows sion of cleared areas and facilitating traditional control and Balciunas, 1997; Buckingham, 2001), but after measures. However, implementation of biological con- discovering that larvae synthesize toxic octapeptides

Figure 1. The strategy employed to control melaleuca in south Florida involved early use of traditional control methods to reduce biomass while biological controls were being developed and implemented.

657 XII International Symposium on Biological Control of Weeds

(Oelrichs et al., 1999), we elected not to release it out guava rust Puccinia psidii G. Winter (Basidiomycetes: of concern over potential negative effects to insectivo- Uredinales), which infects mainly young foliage, ap- rous wildlife. peared during 1997 (Rayachhetry et al., 1997) and is The melaleuca psyllid Boreioglycaspis melaleucae now widespread. Moore was found to be host-specific (Purcell et al., 1997; Wineriter et al., 2003), and was released during The effects of the biological 2002 (Center et al., 2006, 2007). It feeds mainly on the new growth but will also utilize older leaves and the control agents green stems. Furthermore, it completes its life cycle en- Numerous studies aimed at determining the impacts tirely on the plant so it is less restricted by habitat. The of O. vitiosa and B. melaleucae have been conducted or tube-dwelling pyralid Poliopaschia lithochlora (Lower) are ongoing. However, determinations of the individual was highly rated because of its ability to damage mela- effects of one have been confounded by the presence of leuca and its preference for low-lying, humid habitats the other, as well as by the presence of the adventive rust (Galway and Purcell, 2005), but its use of an ornamental fungus and scale insect. These studies have included species, Melaleuca viminalis (Sol. ex Gaertner) Byrnes, comparisons of melaleuca stands with and without the during testing diminished its prospects (M. Purcell, un- agents, caging studies, defoliation experiments, insec- published data). A fergusoninid gall fly, Fergusonina ticide and fungicide exclusion experiments, and before turneri Taylor, and its mutualistic nematode Fergusobia and after comparisons of stand dynamics. melaleucae Davies and Giblin-Davis, also proved to be highly specific (Giblin-Daviset al., 2001) and were first released during 2005 (Blackwood et al., 2006). It has Flower and seed production proven difficult to establish but efforts are continuing. The effects of herbivory by O. vitiosa on melaleuca Most recently a stem-galling cecidomyiid, Lophodiplo- performance were possible early during the release sis trifida Gagné, has proven to be host-specific (S. Win- program when none of the other organisms were pres- eriter et al., unpublished data) and should gain approval ent. Pratt et al. (2005) compared flowering frequency in for release. A bud-feeding weevil Haplonyx multicolor melaleuca stands where the weevil had been released to Lea and a leaf-galling cecidomyiid Lophodiplosis in- stands without them. They found that the likelihood of dentata Gagné are currently under consideration. flowering increased with tree size but that undamaged Two adventive organisms have also recently infested trees were 36 times more likely to reproduce than dam- melaleuca trees in Florida. A pestiferous, undescribed aged trees in similar habitats (Figure 2). Overall, about scale insect (Pemberton, personal communication) was 45% of the weevil-free trees were flowering compared detected in Florida during 1999. It attacks melaleuca to about 2% of infested trees.*** trees as well as some 300 other plant species (Pem- In another study, Pratt et al. (2005) enclosed the berton, 2003; R. Pemberton, unpublished data). The canopies of small (2.9 cm diameter at breast height or

1.0 Damaged trees g Undamaged trees

erin 0.8 ow

0.6

of Trees Fl Trees of 0.4 tion tion 0.2 Propor

0.0 0 1 2 3 4 5 6 Diameter at Breast Height (cm)

Figure 2. Release of the weevil Oxyops vitiosa profoundly affected flowering of melaleuca trees. The proportion of the trees that produced flowers was much lower after being damaged by the weevils regardless of size.

658 Biological control of Melaleuca quinquenervia: goal-based assessment of success dbh) trees with sleeve cages and introduced weevil lar- tiosa, the percentage of embryonic seeds decreased, as vae into the enclosures, either once or twice, to produce did seed viability and germination ability. Seed viabili- one or two defoliations of the young foliage. The sec- ty and germination tests (Van et al., 2005) also revealed ond defoliation was done about 10 weeks after the first. reductions in both measures of seeds from herbivore- These treatments were compared to controls with no attacked trees compared to controls. defoliation or to trees artificially defoliated by manually removing all foliage. Flower production on all trees was monitored monthly for 1 year. The control trees Seedling survival flowered normally during this period, whereas trees ar- Franks et al. (2006) described the effects of the wee- tificially defoliated failed to produce any flowers. Trees vil larvae and the psyllids, alone and in combination, defoliated once or twice by the weevil larvae produced on growth and survival of melaleuca seedlings by cag- a few flowers but numbers were not statistically- dif ing the insects on 26 cm-tall seedlings in field plots. ferent from each other or from the artificial defoliation They compared these results to a natural infestation of treatment (Figure 3). the insects on nearby seedlings. O. vitiosa larvae had Interestingly, in a comparison of ten herbivore- no effect on seedling height, leaf number, or survival, impacted trees with ten non-impacted trees at simi- whereas psyllids caused all of these measures to de- lar, nearby sites at Estero, Florida, Rayamajhi et al. crease by about 55–60% over the 5-month term of the (unpublished data) found that herbivory by O. vitiosa study. About 95% of seedlings survived when protected resulted in higher rates of capsule abortion when com- from psyllids as compared to only 40% when exposed pared to sites without natural enemies. Mean number of to herbivory (Center et al., 2007). capsules in herbivore-impacted infructescences was re- In another study, Tipping et al. (unpublished data) duced by nearly 50% compared to the herbivore-absent found that after becoming infested by both the psyllid site. This decreased density of capsules was apparent as and the weevil, melaleuca trees recruited a much lower gaps in the capsule clusters caused by abortion of the density of seedlings than trees without either herbi- undeveloped fruits. The herbivore-impacted trees were vore. They also compared densities of saplings in plots very similar to those near Brisbane, Australia where the 9 m2 that were periodically treated with insecticide to average infructescence was 5.7 cm long but contained exclude herbivory to saplings in untreated plots. The only 18 capsules (Rayamajhi et al., 2002). The average plots were located in an area that had burned during numbers of seeds per capsule were similar in both the June 1998, resulting in a massive seed rain and thick- Florida and Australian sites. ets of about 1000 seedlings/m2. By the time the study Rayamajhi et al. (unpublished data) have also found was initiated during 2002, these had become saplings that when the trees were subjected to attack by O. vi- and had grown to about 70 cm in height. Densities in

14 A )

o. 12 (n 10 Tree Tree

per per 8

es 6 nc ce

es 4 or fl B In 2 B B 0 Control Herbivory 1 Herbivory 2 Mechanical Damage Treatment

Figure 3. Small trees were caged then subjected to herbivory by Oxyops vitiosa either once (Herbivory 1) or twice (Herbivory 2) or to mechanical defoliation and compared to undefoliated controls. Defoliated trees, regardless of the manner of defoliation, produced very few flowers relative to the controls.

659 XII International Symposium on Biological Control of Weeds the protected plots were virtually unchanged during the Stand dynamics 5-year period of the study as compared to those in the Rayamajhi et al. (2007) studied the dynamics of me- unprotected plots which declined by almost half. laleuca stands before and after the widespread impacts of the biological control agents. They found that the Sapling growth average density of the trees in mature stands declined Tipping et al. (unpublished data) conducted two in- by 72% overall from 15,800 trees/ha during 1996 to secticide exclusion studies on the growth of melaleuca 4400 trees/ha during 2003. Interestingly, the standing saplings in common garden experiments over about biomass based on harvesting studies increased some- a 3-year period. The first experiment investigated the what from an initial average of 263 t/ha to 274 t/ha effect of the melaleuca weevil, O. vitiosa, and supple- during the latter harvest. This was because most of the mental irrigation on the growth of small trees. The second mortality was among the smaller suppressed trees in examined the effects of herbivorous insects (both the understory that represented a small proportion of the psyllids and the weevils) and plant chemotype the biomass. The density of small trees, those with a (nerolidol or viridifloral). In both cases, plants treated dbh of less than 10 cm, decreased 83% from 12,600 to with insecticide more than doubled in stature, where- 2200 trees/ha; density of intermediate-sized trees with as those not treated grew very little. In the first study, a dbh of 10–20 cm decreased 46% from 2600 to 1400 plants attacked by O. vitiosa grew at a much slower trees/ha; density of large trees >20 cm increased from rate compared to the protected plants (Figure 4). The 600 to 800 trees/ha. unprotected plants produced more stem tips per unit of Another study (Rayamajhi et al., 2007) showed that height, creating a shorter, bushier habit, which provided densities decreased between 1997 and 2006, in part due more resource for the tip-feeding insects. Supplemental to self-thinning. The decline accelerated after the ef- irrigation improved the growth of insecticide-treated fects of biological control became apparent and the rate trees but had no effect on trees that were not treated of decline was inversely related to the position of the with insecticide. Chemotype had no apparent effect on trees within the stands. Densities of trees at the periph- the impact of the insects. Seed capsule production was ery, which consisted mostly of small individuals, de- much lower among unprotected plants in both studies. creased by about 6076 individuals/ha/year before 2001

140 Insecticide, Rainfall Only Insecticide, Rainfall + Irrigation 120 No insecticide, Rainfall Only

) No insecticide, Rainfall + Irrigation 100

60 ange in Height (% Height in ange 40 Ch

1 2 2 2 2 2 2 3 3 3 3 3 1 0 2 2 02 2 3 03 -0 -0 - -0 -0 -0 t-02 -0 -0 -0 -0 p- ec ar pr ul ec an un-03 ct-0 -A -J Aug- Oc -J -Apr -Jul-0 -O -Nov-D -Feb-0-M -May-Jun-0 - 1- -Nov -Mar-0 4-Se 29 18 15-Jan-019 13 17 21 25 30 27 15 17-D15 26-Feb-025 30 19-J 31 10 Sample Date

Figure 4. Small trees grown in a common garden plot were treated with insecticide to exclude herbivorous insects and given supplemental irrigation and compared to unprotected and unwatered trees. Protected trees grew vigorously and those receiving supplemental irrigation grew the most. Unprotected trees grew very little and the supplemental irrigation seemed to have little effect.

660 Biological control of Melaleuca quinquenervia: goal-based assessment of success as compared to 16,725 individuals/ha/year after 2001. Stump regrowth Densities in the inner portions of the stands, which con- The ability of melaleuca to sprout from cut stumps tained higher proportions of larger trees, decreased at complicates control. This requires follow-up herbicide relatively constant rates. This further demonstrated the treatment to prevent coppicing and stand regeneration. greater effect of herbivory on smaller trees. The average Several studies have indicated that the flush of foliage diameter of the trees increased, not because they grew associated with this regrowth is highly attractive to both but because of selective mortality of smaller individu- psyllids and weevils, as well as to the rust fungus. Pratt als. This was corroborated by a decrease in or leveling et al. (unpublished data) found that insecticide exclu- off of total basal area coverage during the post-release sion of biological control agents led to an increase in period in contrast to a prior increasing trend. leaf and stem biomass compared to unprotected stumps Despite the finding that the surviving larger domi- (Figure 6). Chronic attack over an 18-month period led nant trees accounted for most of the biomass, biomass to mortality of almost half the unprotected plants. In allocation changed due to extensive defoliation of all a similar insecticide- and fungicide-exclusion study, of the trees. The foliage of large trees growing in dense Rayamajhi et al. (unpublished data) found that the rust stands was limited to the upper branches at the treetops fungus, P. psidii, played an important additive role. and this accounted for only 5.1% of the total biomass Proportion of photosynthetic tissues and the mortality during 1996, before insect-induced defoliation. This de- of stems were higher in treatments involving both increased from 17 to 8 t/ha to represent only 1.5% of the sects (O. vitiosa and B. melaleucae) and rust (P. psidii) total biomass during 2003 (Figure 5). The biomass allo- together than in treatments using either alone. Death of cated to seed capsules decreased by 85% from 6.7 t/ha, the regrowth often led to the death of the stump itself or 0.46% of the total biomass to 1.0 t/ha, or 0.29% of (M. Rayamajhi et al., unpublished data). These data the total biomass. indicate that biological control can compliment, and Litter-traps were placed under mature melaleuca in some cases replace, the use of herbicides for stump stands to collect leaf litter in an attempt to measure the treatment. activity of the biological control agents in the canopy of taller trees. The proportion of fallen leaves that exhib- ited weevil damage symptoms was analysed. Though Discussion the weevil releases began in 1997, the first weevil-damaged leaves did not appear in the traps until 1999 (rep- Clearly, many melaleuca stands have undergone sig- resented by 5% of the trapped leaves) and by 2005, the nificant declines and remaining trees are now in poor proportion of damaged leaves reached approximately condition. However, vast stands of melaleuca still exist 45% (Rayamajhi et al., 2007). This increased percent- that overtly appear unchanged. Yet after closer scrutiny, age of damaged leaves reflected the decreasing propor- we have revealed that the dynamics of these stands tions of leaf biomass (stem to leaf biomass), increasing have changed in very significant ways. Fewer trees tree mortality, and decreasing tree densities. now produce flowers, those that do flower produce

Figure 5. The proportion of the total tree biomass allocated to foliage declined dramatically due to defoliation primarily by Oxyops vitiosa and the psyllid Boreiogly- caspis melaleucae.

661 XII International Symposium on Biological Control of Weeds

Figure 6. Regrowth from stumps, reflected by the biomass of stems and leaves produced after felling of the original trees, was substantially more when regrowth was treated with insecticide thus reducing the effects of the weevils and the psyllids. fewer inflorescences and the inflorescences produced program for the tremendous support provided by the contain fewer individual blossoms. Many of the fruits many conservation interns that have been involved in abort and those that do manage to set seed produce a this program. smaller proportion of viable seeds. The constant defoliation of the stem tips causes the capsules to desiccate and release seeds during drier periods when conditions References are unfavourable for germination. Those that do fall, Balciunas, J.K. (1990) Australian insects to control melaleuca. lodge in a favourable site and manage to germinate are Aquatics 12, 15–19. infested by psyllids that kill a large proportion before Balciunas, J.K., Bowman, G.J. and Edwards, E.D. (1993a) they attain a significant size. If they survive, they grow Herbivorous insects associated with paperbark Melaleuca slowly due to constant defoliation and produce few quinquenervia and its allies: I. Noctuoidea (Lepidoptera). flowers. Meanwhile, existing stands have nearly been Australian Entomologist 20, 13–24. removed from publicly held lands and those on private Balciunas, J.K., Burrows, D.W. and Edwards, E.D. (1993b) lands are less invasive. Hence, the goals of the project, Herbivorous insects associated with the paperbark tree as stated above, have been met so the project should Melaleuca quinquenervia and its allies: II. Geometridae be considered a success. It is not yet a ‘complete’ suc- (Lepidoptera). Australian Entomologist 20, 91–98. cess in that biological control is more effective in some Balciunas, J.K., Burrows, D.W. and Purcell, M.F. (1994) In- sects to control melaleuca I: Status of research in Austra- habitats and during some periods than others but addi- lia. Aquatics 16, 10–13. tional agents that are currently under development may Balciunas, J.K., Burrows, D.W. and Horak, M. (1995a) fill these gaps. Herbivorous insects associated with the paperbark tree Melaleuca quinquenervia and its allies: IV. Tortricidae Acknowledgements (Lepidoptera). Australian Entomologist 22, 125–135. Balciunas, J.K., Burrows, D.W. and Purcell, M.F. (1995b) The research reported herein was supported by funding Insects to control melaleuca II: Prospects for additional from the South Florida Water Management District, agents from Australia. Aquatics 17, 16 and 18–21. the Florida Department of Environmental Protection, Balciunas, J.K., Burrows, D.W. and Purcell, M.F. (1995c) the US Army Corps of Engineers, the Miami-Dade Australian insects for the biological control of the pa- County Department of Environmental Resource Man- perbark tree, Melaleuca quinquenervia, a serious pest of Florida, USA, wetlands. In: Delfossee, E.S. and Scott, agement, and Lee County as well as by the USDA- R.R. (eds) Proceedings of the Eighth International Sym- Agricultural Research Service Areawide Projects. We posium on Biological Control of Weeds. DSIR/CSIRO, thank all past and present staff of the USDA–ARS Aus- Melbourne, Australia, pp. 247–267. tralian Biological Control Laboratory and the Invasive Blackwood, S., Pratt, P.D. and Giblin-Davis, R. (2006) Bud- Plant Research Laboratory. We are also indebted to gall fly release for biocontrol of Melaleuca in Florida. the Student Conservation Service and the AmeriCorps Biocontrol News and Information 26(2), 48N.

662 Biological control of Melaleuca quinquenervia: goal-based assessment of success

Bodle, M.J., Ferriter, A.P. and Thayer, D.D. (1994) The bio Crawley, M.J. (1989b) The successes and failures of weed logy, distribution, and ecological consequences of Mela- biocontrol using insects. Biocontrol News and Informa- leuca quinquenervia in the Everglades. In: Davis, S.M tion 10, 213–223. and Ogden, J.C. (eds) Everglades—The Ecosystem and Di Stefano, J.F. and Fisher, R.F. (1983) Invasion potential of its Restoration. St. Lucie Press, Delray Beach, FL, USA, Melaleuca quinquenervia in Southern Florida, USA. For- pp. 341–355. est Ecology and Management 7, 133–141. Buckingham, G.R. (2001) Quarantine host range studies with Dray, F.A., Bennett, B.C. and Center, T.D. (2006) Invasion Lophyrotoma zonalis, an Australian sawfly of interest for history of Melaleuca quinquenervia (Cav.) S.T. Blake in biological control of melaleuca, Melaleuca quinquenervia, Florida. Castanea 71, 210–225. in Florida. BioControl 46, 363–386. Franks, S.J., Kral, A.M. and Pratt, P.D. (2006) Herbivory by in- Burrows, D.W. and Balciunas, J.K. (1997) Biology, distribu- troduced insects reduces growth and survival of Melaleuca tion and host range of the sawfly, Lophyrotoma zonalis quinquenervia seedlings. Environmental Entomology 35, (Hym., Pergidae), a potential biological control agent for 366–372. the paperbark tree, Melaleuca quinquenervia. Entomoph- Galway, K.E. and Purcell, M.F. (2005) Laboratory life his- aga 42, 299–313. tory and field observations of Poliopaschia lithochlora Burrows, D.W. and Balciunas, J.K. (1999) Host-range and (Lower) (Lepidoptera: Pyralidae), a potential biological distribution of Eucerocoris suspectus (Hemiptera: Miri- control agent for Melaleuca quinquenervia (Myrtaceae). dae), a potential biological control agents for the paper- Australian Journal of Entomology 44, 77–82. bark tree Melaleuca quinquenervia (Myrtaceae). Environ- Giblin-Davis, R.M., Makinson, J., Center, B.J., Davies, K.A., mental Entomology 28, 290–299. Purcell, M., Taylor, G.S., Scheffer, S.J., Goolsby, J. and Burrows, D.W., Balciunas, J.K. and Edwards, E.D. (1994) Center, T.D. (2001) Fergusobia/Fergusonina-induced Herbivorous insects associated with the paperbark tree shoot bud gall development on Melaleuca quinquenervia. Melaleuca quinquenervia and its allies: III. Gelechioidea Journal of Nematology 33, 239–247. (Lepidoptera). Australian Entomologist 21, 137–142. Hoffmann, J.H. (1995) Biological control of weeds: the way Burrows, D.W., Balciunas, J.K. and Edwards, E.D. (1996) forward, a South African perspective. In: British Crop Pro- Herbivorous insects associated with the paperbark tree tection Council Proceedings No. 64: Weeds in a Changing Melaleuca quinquenervia and its allies: V. Pyralidae (Lep- World. BCPC, Farnham, Surrey, UK, pp. 77–89. idoptera). Australian Entomologist 23, 7–16. Laroche, F. and Ferriter, A.P. (1992) The rate of expansion Center, T.D., Van, T.K., Rayachhetry, M., Buckingham, G.R., of melaleuca in south Florida. Journal of Aquatic Plant Dray, F.A., Wineriter, S.A., Purcell, M.F. and Pratt, P.D. Management 30, 62–65. (2000) Field colonization of the melaleuca snout beetle Louda, S. M. and Stiling, P. (2004) The double-edged sword (Oxyops vitiosa) in south Florida. Biological Control 19, of biological control in conservation and restoration. Con- 112–123. servation Biology 18, 50–53. Center, T.D., Van, T.K., Dray, F.A., Franks, S.J., Rebelo, M.T., Meskimen, G.F. (1962) A Silvical Study of the Melaleuca Pratt, P.D. and Rayamajhi, M.B. (2005) Herbivory alters Tree in South Florida. MS thesis. University of Florida, competitive interactions between two invasive aquatic Gainesville, FL, USA, 177 pp. plants. Biological Control 33, 173–185. McFadyen, R.E.C. (1998) Biological control of weeds. An- Center, T.D., Pratt, P.D., Tipping, P.W., Rayamajhi, M.B., nual Review of Entomology 43, 369–393. Van, T.K., Wineriter, S.A., Dray, F.A., Jr. and Purcell, McFadyen, R.E.C. (2000) Successes in biological control M. (2006) Field colonization, population growth, and of weeds. In: Spencer, N.R. (ed.) Proceedings of the dispersal of Boreioglycaspis melaleucae Moore, a bio- X International Symposium on Biological Control of logical control agent of the invasive tree Melaleuca Weeds. Montana State University, Bozeman, MT, USA., quinquenervia (Cav.) Blake. Biological Control 39, pp. 3–14. 363–374. Oelrichs, P.B., MacLeod, J.K., Seawright, A.A., Moore, Center, T.D., Pratt, P.D., Tipping, P.W., Rayamajhi, M.B., M.R., Ng, J.C., Dutra, F., Riet-Correa, F., Mendez, M.C. Van, T.K., Wineriter, S.A. and Dray, F.A., Jr. (2007) Ini- and Thamsborg, S.M. (1999) Unique toxic peptides iso- tial impacts and field validation of host range forBoreio - lated from sawfly larvae in three continents. Toxicon 37, glycaspis melaleucae Moore (Hemiptera: Psyllidae), a 537–544. biological control agent of the invasive tree Melaleuca Ogden, J.C. (2005) Everglades ridge and slough conceptual quinquenervia (Cav.) Blake (Myrtales: Myrtaceae: Lep- ecological model. Wetlands 25, 810–820. tospermoideae). Environmental Entomology 36, 569– Pemberton, R.W. (2003) Potential for the biological control 576. of the lobate lac scale, Paratachardina lobata lobata (He- Coetzee, J.A., Center, T.D., Byrne, M.J. and Hill, M.P. (2005) miptera: Kerridae). Florida Entomologist 86, 353–360. Impact of the biocontrol agent Eccritotarsus catarinensis, Peschken, D.P. and McClay, A.S. (1995) Picking the target: a sap-feeding mirid, on the competitive performance of A revision of McClay’s scoring system to determine the waterhyacinth, Eichhornia crassipes. Biological Control suitability of a weed for classical biological control. In: 32, 90–96. Delfosse, E.S. and Scott, R.R. (eds) Proceedings of the Crawley, M.J. (1989a) Plant life-history and the success of Eighth International Symposium on Biological Control of weed biological control projects. In: Delfosse, E.S. (ed.) Weeds. CSIRO, Melbourne, Australia, pp. 137–143. Proceedings of the VII International Symposium on Bio- Pratt, P.D., Rayamajhi, M.B., Van, T.K., Center, T.D. and Tip- logical Control of Weeds. Instituto Sperimentale per la ping, P.W. (2005) Herbivory alters resource allocation and Patologia Vegetale, Ministero dell’ Agricoltura e delle Fo- compensation in the invasive tree Melaleuca quinque- reste, Rome, Italy, pp. 17–26. nervia. Ecological Entomology 30, 316–326.

663 XII International Symposium on Biological Control of Weeds

Purcell, M.F. and Balciunas, J.K. (1994) Life history and dis- forests in Florida: analyses of natural-enemy impacts on tribution of the Australian weevil Oxyops vitiosa (Coleop- stand dynamics. Plant Ecology 192, 119–132. tera: Curculionidae), a potential biological control agent Tomlinson, P.B. (1980) The Biology of Trees Native to Tropi- for Melaleuca quinquenervia (Myrtaceae). Annals of the cal Florida. Harvard University Printing Office, Allston, Entomological Society of America 86, 867–873. MA, USA, 480 pp. Purcell, M.F., Balciunas, J.K. and Jones, P. (1997) Biology Van, T.K., Rayachhetry, M.B. and Center, T.D. (2000). Esti- and host-range of Boreioglycaspis melaleucae (Hemip- mating above-ground biomass of Melaleuca quinquenervia tera: Psyllidae), a potential biological control agent for in Florida, USA. Journal of Aquatic Plant Management Melaleuca quinquenervia (Myrtaceae). Environmental 38, 62–67. Entomology 26, 366–372. Van, T.K., Rayamajhi, M.B. and Center, T.D. (2005) Seed Rayachhetry, M.B., Elliott, M.L. and Van, T.K. (1997) Natu- longevity of Melaleuca quinquenervia: a burial experi- ral epiphytotic of the rust Puccinia psidii on Melaleuca ment in south Florida. Journal of Aquatic Plant Manage- quinquenervia in Florida. Plant Disease. 81, 831. ment 43, 39–42. Rayachhetry, M.B., Elliott, M.L. and Van, T.K. (1998) Re- White, P. (1994) Synthesis: vegetation pattern and process in generation potential of the canopy-held seeds of Melaleu- the Everglades ecosystems. In: Davis, S.M. and Ogden, ca quinquenervia in south Florida. International Journal J.C. (eds) Everglades, The Ecosystem and Its Restoration. of Plant Science 159, 648–654. St. Lucie Press, Florida, USA, pp. 445–458. Rayachhetry, M.B., Van, T.K., Center, T.D. and Laroche, F. Wineriter, S.A., Buckingham, G.R. and Frank, J.H. (2003) (2001) Dry weight estimation of the aboveground com- Host range of Boreioglycaspis melaleucae Moore (He- ponents of Melaleuca quinquenervia trees in southern miptera: Psyllidae), a potential biological control agent of Florida. Forest Ecology and Management 142, 281–290. Melaleuca quinquenervia (Cav.) S.T. Blake (Myrtaceae), Rayamajhi, M.B., Van, T.K., Center, T.D., Goolsby, J.A., under quarantine. Biological Control 27, 273–292. Pratt, P.D. and Racelis, A. (2002) Biological attributes of Woodall, S.L. (1981) Integrated methods for melaleuca con- the canopy held melaleuca seeds in Australia and Florida, trol. In: Geiger, R.K. (ed.) Proceedings of Melaleuca US. Journal of Aquatic Plant Management 40, 87–91. Symposium, Sept. 23–24 1980. Florida Department of Rayamajhi, M.B., Van, T.K., Pratt, P.D., Center T.D. and Tip- Agriculture and Consumer Services, Division of Forestry, ping, P.W. (2007). Melaleuca quinquenervia dominated Gainesville, FL, USA, pp. 135–140.

664