100 Plant Protection Quarterly Vol.24(3) 2009 ca. 161, range: 115–189 mg ‘wet weight’ of seeds collected from pods on trees and stored) as in Australia (mean ca. 85, range: 56–103 mg) (Woods 1988). Also, P. aculeata Review in Argentina has been reported anecdo- tally to be morphologically different from Australian plants, having brown, cylindri- cal pods (H. Cordo personal communica- tion 2006) and distinctive plant architec- ture (M. Julien personal communication The biology of Australian weeds 2006). 54. Parkinsonia aculeata L. Description The following description is based on Rieks Dekker van KlinkenAD, Shane D. CampbellBD, Tim A. HeardAD, John our own observations and those of others McKenzieB and Nathan MarchC (Dimmit 1987, Stewart et al. 1992, Vitelli A CSIRO Entomology, Long Pocket Laboratory, 120 Meiers Road, 1995, Ross 1998, Hawkins 2001, Parsons Indooroopilly, Queensland 4058, Australia. and Cuthbertson 2001). In Australia P. aculeata is an unmistake- B Queensland Department of Primary Industries and Fisheries, Tropical Weeds able plant with smooth, green bark and Research Centre, PO Box 187, Charters Towers, Queensland 4820, Australia. very distinctive pendulous leaves with C Queensland Department of Primary Industries and Fisheries, PO Box 53, minute, easily-shed pinnules (Figure 1). Cloncurry, Queensland 4824, Australia. Plants typically grow to 5 to 7 m high DCRC for Australian Weed Management. and wide, although they can grow to 20 m high, with trunks to 40 cm in diameter at breast height at maturity (MacDicken and Brewbaker 1984). Plants can be either single-stemmed or multi-stemmed, with Name Common name steeply ascending branches. Roots typi- Botanical name The plant is usually referred to as par- cally have a shallow main axis and nu- The genus Parkinsonia (Caesalpinioideae) is kinsonia in Australia and Mexican palo merous surface laterals. Leaves are shortly named after John Parkinson, a seventeenth verde and retama in the American litera- stalked, bipinnate, each with 1–4 pairs of century English apothecary and botanist. ture. However, overseas it has many local pinnae. The pinnular rachi are very thin, The species name aculeata is derived from names, including Jerusalem thorn, blue green, fl attened and ridged, and are 20 to the Latin word meaning ‘prickly’ (Wilson palo verde, horse bean tree, sessaban and 40 cm long with numerous minute (4–10 and Miller 1987, Stewart et al. 1992). The Barbados fl ower fence (Hawkins 2001). mm long) pinnules (secondary leafl ets) on plant was fi rst described in Charles Plum- each edge (Figure 2a). Stiff, needle-sharp ier’s Nova Plantarum Americanum Gen- Taxonomy spines (modifi ed leaf rachises) 5–15 mm era (1703) from plants growing in the West Parkinsonia aculeata is easily delimited long are formed at the base of each leaf Indies (Hawkins 2001). morphologically from all other Parkin- (Figure 2a). The spines persist on the older Relationships between caesalpinioid sonia species (Hawkins 2001); however, branches and stems. genera are still under dispute, but genera considerable intra-specifi c genetic varia- Flowers are bright yellow, fragrant, 2 that are closely related to Parkinsonia in- tion occurs across its distribution. Several cm in diameter with fi ve-petalled corol- clude Delonix, Peltophorum and Lemuropis- genetically distinct populations have been las (Figure 2b). They are bilaterally sym- um (Doyle et al. 2000, Herendeen et al. 2003, identifi ed across the Americas (Hawkins et metrical, with four petals 6–15 mm long Haston et al. 2005, Hawkins et al. 2007). al. 2007): (1) northern and western Mexico, becoming refl exed, and one erect petal The genus Parkinsonia is considered to south-western USA and Cuba; (2) eastern which is either completely orange or has be congeneric with the paraphyletic Cen- and southern Mexico and south-eastern orange spots. Each fl ower is on a slender tral American genus, Cercidium (Hawkins USA; (3) Venezuela; (4) Central America; stalk and fl owers are borne in groups of 8 et al. 2007). All species are small trees and and (5) Argentina. The Argentine lineage to 12 in loose racemes (Figure 2b) arising spreading shrubs that are found mainly (5) is estimated to have diverged from oth- from the leaf axils. in subtropical and tropical arid habitats er lineages (1–4) about 9.1 mil- (Polhill and Vidal 1981, Hawkins 1996, lion years ago, and the north- 1997, Hawkins and Harris 1998, Hawkins ern Mexico lineage (1) from et al. 1999, Haston et al. 2005). Most species the Mesoamerican-Venezuelan are American, distributed from the south- lineages (2–4) about 5.2 million ern USA to Argentina (P. aculeata, P. fl orida years ago (both pre-dating (Benth. ex A.Gray) S.Watson, P. microphyl- formation of the Isthmus of la Torr., P. peruviana C.E.Hughes, Daza & Panama). Additional diver- Hawkins, P. praecox (Ruiz Lopez & Pavon) gent populations may exist in Hawkins, P. texana (A.Gray) S.Watson South America, but these have and C. andicola Griseb). In addition, there not been sampled genetically. are three species in east Africa (P. scioana Intra-specifi c morphologi- (Chiovenda) Brenan, P. raimondoi Brenan cal variation in P. aculeata has and P. anacantha Brenan), one in southern not been studied systemati- Africa (P. africana Sond.) and two unde- cally, including in relation to scribed species from the Horn of Africa genotypes. However, seeds (Polhill and Vidal 1981). have been reported to be twice as heavy in Arizona (mean Figure 1. An adult Parkinsonia aculeata plant. Plant Protection Quarterly Vol.24(3) 2009 101 Pods, each containing 1–11 (typically 1–3) seeds, are 3 to 13 cm long, about 7 (a) mm wide, fl attened, constricted between the widely-spaced seeds, tapered at both ends and straw coloured when ripe (Fig- ure 2c). Seeds are oblong, 8–12 mm long and 4–7 mm wide. The testa is green and soft when immature, becoming mottled brown and very hard when mature. The cotyledons are light yellow in the mature seed and the endosperm is tough, plastic and translucent, becoming brittle on ma- turity (Woods 1988). The embryo accounts for about 27–34% of total seed weight, the endosperm 25–33%, and the seed coat 40% (Woods 1988). Parkinsonia aculeata has a mitotic chro- mosome number of 2n = 28 (Hawkins et al. 1999).

Distinguishing Parkinsonia aculeata from other taxa In Australia P. aculeata can be confused with prickly acacia (Acacia nilotica (L.) Del.), mi- mosa bush (A. farnesiana (L.) Willd.) and mesquite (Prosopis species), with which it frequently co-occurs. All four taxa pro- (b) duce spines, yellow fl owers and bean-like seed pods, and are sometimes collectively referred to as ‘prickle bushes’. P. aculeata can be most readily distinguished by its characteristic leaf structure, photosynthet- ic bark and large fl owers that have open petals, as opposed to the other prickle bushes that have fl owers with minute pet- als and form tight clusters (pompoms or lambs tails).

History Parkinsonia aculeata was being cultivated in greenhouses in England by 1739. At that time it was very common in the Spanish West Indies and had been introduced into the English settlements in North America, for the ‘beauty and sweetness of its fl ow- ers’, and was predicted to ‘soon become common in all host countries’ (Miller 1739). Since then it has been introduced actively around the world, often fi rst be- ing introduced as an ornamental tree in towns and botanic gardens (Hughes 1989), Figure 2. Parkinsonia aculeata plant but also for use as a hedging plant and parts: (a) leaf (long strap-like rachi (c) for shade, windbreak, soil nitrogen fi xa- with small pinnules along their tion and rehabilitation purposes. These edges) and thorns, (b) fl owers and introductions have generally been casual, (c) pods. unsystematic and from unknown sources (Hughes 1989). Some details of its intro- duction through the British Empire, and later the British Commonwealth, have been reported by Streets (1962) and Troup evergreen hedge, was being advocated as and Joshi (1983), and more generally in early as 1876 (von Mueller 1876). The fi rst Hawkins 2001. The Kew Herbarium holds Australian record we could fi nd was of a specimens collected prior to 1850 from Iran planting in the Botanic Gardens in Syd- (formerly Persia), Senegal, India, Pakistan ney prior to 1901. However, its invasive- and the Mascarenes (Mauritius, Réunion ness in Central Queensland by the early and Rodrigues) (Hawkins 2001). 1900s (Bailey 1906) suggests that it was Knowledge of its introduction and introduced soon after this area was widely spread in Australia is based largely on settled in the 1860s. Genetic evidence sug- anecdote. Its planting in Australia, as an gests Australian populations may have 102 Plant Protection Quarterly Vol.24(3) 2009 originated from Venezuela and/or other Central and South America (Hawkins et ornamental, hedging, fodder and shade closely-related Mesoamerican populations al. 2007) (Figure 3). tree (Stewart et al. 1992, Woods 1992, that were not included in the sampling In the United States its distribution in- Hawkins and Harris 1998, Wagner et al. (Hawkins et al. 2007). Also, populations cludes the South Texas Plains and Coastal 1999, PIER 2000, Hawkins 2001, R.D. van sampled from Australia and Kenya may Prairies in Texas (Everitt 1983), the Sono- Klinken unpublished) (Figure 4). It fre- have originated from a single transloca- ran Desert (Carter 1974), and the warmer quently naturalizes; however, other than tion event (Hawkins et al. 2007). parts of the Mojave Desert (Dimmit 1987). in Australia, records of it becoming a trou- Anecdotal evidence suggests that, once Native populations in Central America ap- blesome weed are relatively rare. In New in Australia, P. aculeata was quickly spread pear to be quite restricted and disjunct, oc- Caledonia it has been reported as ‘com- as an ornamental or shade tree for plant- curring as a series of isolated dense stands monly occurring as populations which ing around homesteads, bores and dams on deep black vertisols in southern Mexi- signifi cantly affect yield or profi tability, (Anon. 1978, Miller and Pickering 1980, co, Guatemala, Nicaragua and Costa Rica and its removal would have useful ben- Woods 1986, Wilson and Miller 1987). For (Hughes and Styles 1984, Hughes 1989) efi ts’ (Swarbrick 1997). An extensive natu- example, a mature P. aculeata plant was (Figure 3). These stands typically grow on ralized population occurs near Nakuru in growing in front of the Borroloola Police saline, high-pH, seasonally waterlogged Kenya (Hawkins 2001). Station prior to 1930, and is thought to sites that were formerly lake-beds, or are have been the source of seeds for plantings seasonally fl ooded coastal estuaries and Introduced – Australia at the homestead on nearby Mallapunyah lagoons. It does not occur in adjacent dry The distribution of P. aculeata has been Station where it subsequently became in- forest or semi-arid thorn scrubs. In Brazil mapped nationally in a 50 × 50 km grid, vasive over much of the 434 900 ha pas- it occurs in semi-arid regions of the Ser- mainly through existing distributional toral lease (Deveze 2004). P. aculeata was tao (Caatinga vegetation) which occupies records held by state departments and planted at Government Bore on Muckaty most of the Brazilian northeast (Barbosa through expert knowledge (Figure 5). Station (Barkly Tablelands, Queensland) and Prado 1991). When considered at that grid scale, P. ac- around the late 1930s for shade, and was uleata is now estimated to be present on commonly planted for shade throughout Introduced – worldwide over 3.3 million ha of Australia, although the Victoria River District in the 1960s Parkinsonia aculeata has a pan-tropical densities are very low throughout most (Larry Johns personal communication distribution following introduction as an grid cells. 2003). By 1906 it was already considered a weed in parts of Queensland, especially along the Fitzroy River (Bailey 1906). By 1972 it had taken a fi rm hold in the Victo- ria River, Darwin and Gulf Districts and was reported to be undergoing a ‘popu- lation explosion’ despite having been in the Northern Territory for ‘years’ (Anon. 1972). By 1988 it was considered weedy in the Northern Territory (with major infes- tations reported from the Victoria River district and on black clay soil on the Bark- ly Tablelands), Queensland (fl at downs country near Richmond and land adjacent to the Burdekin, Isaacs and Fitzroy Rivers) and Western Australia (along rivers in the Pilbara and Kimberley regions, with im- portant infestations along the Ord, Fitzroy Figure 3. Distribution of Figure 5. Current distribution and and De Grey Rivers) (Woods 1988). It was Parkinsonia aculeata in the abundance of Parkinsonia aculeata identifi ed as one of Australia’s 20 Weeds Americas ({) including dense in Australia. Source: Queensland of National Signifi cance in 1999 following stands in Central America (z). Department of Primary Industries. a national prioritization process (Thorpe and Lynch 2000).

Distribution Native Parkinsonia aculeata is native to the Ameri- cas, where it is widely distributed in hot and dry regions (Figure 3); however, his- torically there has been considerable dis- pute as to its native range (Hawkins et al. 2007). This has probably been due to P. aculeata having been transported inten- tionally in the Americas for many centu- ries, and in much of its range only being recorded from highly disturbed habitats (such as roadsides, urban and wasteland sites), often in association with intentional plantings as ornamentals (Hughes 1989). Recent genetic work has shown that its na- Figure 4. Countries or states (in USA) (shaded) where Parkinsonia aculeata tive distribution includes parts of North, has been recorded (see text). Plant Protection Quarterly Vol.24(3) 2009 103 Most infestations occur across semi-ar- Kimberley and Pilbara regions of Western (a) id and semi-humid Australia, especially in Australia, where scattered infestations oc- central and north Queensland, the Barkly cur along many hundreds of kilometres Region and the Victoria River District of of water courses. In the Kimberley Re- the Northern Territory, and the Kimberley gion infestations are most extensive on and Pilbara Regions of Western Australia. the Fitzroy River. Around Kununurra it Although it is widespread in these regions, is primarily restricted to the banks of the dense patches are associated primarily Ord River, fl ood-outs and on the edges of with fl ood-outs, water infrastructure (such wetlands. Most infestations are scattered, as ‘turkey nests’), watercourses and the and dense thickets are generally isolated. edges of seasonally-fl ooded fresh-water In the Pilbara Region P. aculeata is almost wetlands. Elsewhere in Australia records entirely restricted to the riparian zone of are mostly of isolated plants, or relatively the major rivers (all ephemeral), includ- restricted, scattered infestations. ing along the Fortescue, Maitland, Hard- In Queensland extensive infestations ing, De Grey and Robe Rivers, and in the occur primarily in riparian, fl ood-out and Millstream National Park. Dense stands seasonally flooded habitats in Central occur on the Fortescue and De Grey Riv- (b) Queensland (e.g. the Fitzroy River catch- ers. South of the Pilbara Region P. aculeata ment), upper Lake Eyre Basin and in the occurs as small isolated populations or as Gulf of Carpentaria, with some infesta- individual plants in the Upper Gascoyne tions up to several kilometres across. Ma- and Murchison Districts. A single P. aculea- jor aerial surveys have been undertaken ta plant (about 1 m tall) has been found of the Cooper Creek and Georgina – Dia- naturalized in the Southwest Agricultural mantina catchments by Desert Channels area (J. Peirce personal communication Queensland since 2004. Analysis of these 2001). survey data indicates the total area of in- Naturalized populations of P. aculeata festation across the region to be 600 000 ha have been recorded from four sites in (Brett Carlsson personal communication New South Wales: two clumps of indi- 2008). The bulk of these infestations are of vidual plants near Broken Hill, a 250 ha low density and follow more or less lin- patch along 20 km of the Narran River and ear bands associated with watercourses. ‘about a dozen’ plants at an abandoned While few populations are greater than homestead in Bourke Shire (March 2004, (c) 200 m diameter, some follow rivers semi- in Deveze 2004). The Narran River popu- continuously for up to 80 km. A similar lation is thought to have originated from pattern is seen on the Cape River catch- Yerambah, a now-abandoned homestead ment in central-north Queensland (Lawes upstream of the current infestation. In ad- et al. 2003). With few exceptions, P. aculeata dition, there have been reported cases of populations on Cape York Peninsula and single trees from council areas close to the southern Queensland are very sparse. Queensland border and from the Royal In the Northern Territory P. aculeata ex- Botanical Gardens in Sydney (P. Gray per- tends from the sandy plains of the Tan- sonal communication 2000). ami to the wetlands of the wet-dry trop- Records from South Australia are re- ics; however, it is suffi ciently dense to be stricted to isolated plants in the vicinity considered a problem only in parts of the of Port Augusta, Woomera and Copley Gulf District, Barkly Tablelands and Vic- townships. All occurrences are associated toria River District (Flanagan et al. 1996). with historical plantings although plants Infestations reach their greatest extent have naturalized at each site (N. March in seasonally-flooded black soil plains personal communication 2008). on the Barkly Tablelands, where mono- Figure 6. Potential Australian cultures can reach several kilometres in Potential distribution in Australia distribution of Parkinsonia length. In the Gulf district populations Much of northern and eastern Australia aculeata with existing climate (a), are mostly scattered, although P. aculeata is probably climatically suitable for P. ac- and following a 2°C increase in continues to spread along rivers such as uleata, provided adequate soil moisture is temperature and 10% reduction the MacArthur, Robinson and Calvert. In available, with conditions being optimal the Katherine River Region and the Victo- in central Queensland (Figure 6a). On the in rainfall (b), and 10% increase ria River District it is widely distributed, broad scale P. aculeata has probably natu- in rainfall (c). Modelling was occurring in low densities along most ralized in the majority of suitable catch- conducted in CLIMEX (Lawson et major watercourses, although there are ments (Figure 5). Within catchments P. al. 2008, R.D. van Klinken et al. in some high density patches such as on the aculeata is generally very sparsely and/ prep.). Roper River. In central Australia (Lawes or locally distributed, but there is lit- and Grice 2007) and the wet-dry tropics of tle doubt that P. aculeata will continue to the Northern Territory (the ‘Top End’) P. spread through the wetter habitats within aculeata is widely distributed but sparse, its current range. Special efforts are cur- Climate change is expected to result with thickets rarely exceeding more than rently underway to prevent its spread in a southward extension of highly suit- a dozen trees. Populations in central Aus- into Cape York Peninsula, the Lake Eyre able areas in eastern Australia as a result tralia are mostly within 30 m of water- and Murray Darling basins in Queens- of reduced cold stress (van Klinken et al. courses. land and the blue-bush (Maireana spp.) unpublished) (Figure 6). Also, in south- Parkinsonia aculeata has been recorded swamps in the Barkly Tablelands (Deveze west Australia it is expected that there from many of the river systems in the 2004). will be improved growing conditions and 104 Plant Protection Quarterly Vol.24(3) 2009 reduced cold-wet stress. Reduced rainfall is expected to result in the northern (tropi- (a) cal) interior becoming less suitable (Figure 6b), while increased rainfall is expected to increase the suitability of much of Aus- tralia (Figure 6c). Landscape-scale modelling has been conducted to predict habitat suitability and susceptibility in the Desert Channels Region in south-western Queensland (R.D. van Klinken et al. unpublished data), where P. aculeata is predicted to remain restricted to the most mesic parts of the landscape.

Habitat Parkinsonia aculeata has an outstand- ing ability to survive and grow under a wide range of environmental conditions (Hughes 1989). This includes arid regions to wet-dry tropical regions, with annual rainfall typically ranging between 250 and 1400 mm. It can survive on a wide range of soil types, including deep sand, shallow and skeletal soils, rocky gullies and heavy, black cracking clay soils, where few oth- (b) er woody plants can survive (MacBride 1943, Stewart et al. 1992, Hawkins 2001). It grows well on moderately to extremely sa- line soils (Hughes 1989, Stewart et al. 1992, Hawkins 2001), occurring on salt-affected sandy rises along the Gulf of Carpentaria coastline in both the Burketown (Queens- land) and Borroloola (Northern Territory) areas, and in seasonally-fl ooded coastal estuaries in Central America (Hughes 1989) (Figure 7a). It is also very tolerant to extremes in pH (3–11) (Stewart et al. 1992, Hawkins 2001, Parsons and Cuthbertson 2001). In coastal Peru it grows in arid habi- tats of extremely low rainfall in which few other species will grow (Figure 7b). Parkinsonia aculeata plants can continue to grow with their root-systems inundated by water, and can survive more than nine months with the lower portion of their trunks under water (van Klinken 2005). Figure 7. Native Parkinsonia aculeata populations in contrasting habitats: Prolonged, repeated inundation will, how- ever, result in stress and death. It is mildly (a) dense population on black clay soil in an estuary (salt marsh) near Palo frost tolerant, although dieback occurs Verde National Park, Costa Rica and (b) a prostrate adult growing in the after intense or extended frosts (Hocking coastal desert of Peru. 1993, in Hawkins 2001); for example, a cold frost period of about −6°C resulted in top-kill or death of most P. aculeata plants at study sites in central Australia (R.D. van Barbosa and Prado 1991, Deveze 2004). Growth and development Klinken unpublished data). P. aculeata is Thickets can also form in upland habi- Information on growth and development, not limited by high temperatures in Aus- tats that are never inundated, especially reproduction and population character- tralia. in semi-arid and semi-humid regions istics has largely been obtained through Thicket formation rarely occurs away (Everitt 1983, van Klinken 2005, unpub- research conducted in Australia between from the most mesic parts of the land- lished data). However, the occurrence of 1999 and 2007. Much of this work is yet to scape, being most common along rivers, P. aculeata in upland habitats may, in part, be published. adjacent to water-storages, and in season- be seed-limited as a result of poor seed dis- ally fl ooded habitats such as river levees, persal in the absence of fl ooding (see later). Morphology depressions that receive run-on or fl ood- Thicket formation is generally restricted to Adult trees are typically shallow-rooted water, the margins of lagoons and periodi- areas that are heavily grazed by livestock with numerous surface laterals (Parsons cally inundated black soil plains (Wiggins or feral animals, provided the herbivores and Cuthbertson 2001). Trees can re- and Porter 1971, Scrifres 1980, Benson and (e.g. macropods, goats or sheep) do not shoot laterally from the main trunk down Darrow 1981, Everitt 1983, Woods 1988, also browse P. aculeata plants. to about 20 cm below the soil surface Plant Protection Quarterly Vol.24(3) 2009 105 following damage, such as after fi re, frost, seedlings survived drought treatments (no Reproduction prolonged flooding, termite-feeding or drought, 30% and 10% above wilting point Floral biology mechanical disturbance. This can result and no water) imposed from 4–18 weeks Flowers of P. aculeata are bee pollinated in 20 or more branches arising from the of age in sand, loam and clay soils (Lawes (Eisikowitch and Masad 1982, Lewis et al. root crown. Reshooting also occurs (typi- 2004). Above ground biomass increased 2000). Generalist bees such as species in cally along the whole main stem) if the in all soil types, and growth was fastest in the genera Canephorula, Centris, Colletes tree has been pushed over by floods. clay, relatively rapid in loam, but restricted (references in Lewis et al. 2000) are record- Above-ground biomass of juveniles and in the sandy soils. ed as fl ower visitors in the native range adults can be estimated using plant height, and large carpenter bees (Xylocopa sp.) canopy diameter or stem cross-sectional Phenology have been observed on P. aculeata fl owers area, although stem area is the most robust Leaf production occurs throughout the in Australia and Costa Rica (T.A. Heard predictor (Grice et al. 2002). The relation- year, provided conditions remain warm unpublished data). The UV-absorbent fl ag ships between size and biomass are linear and trees remain healthy; however, a petal of P. aculeata has been shown to be when plotted on log-log scales. Plant size major fl ush of leaf production generally an important recognition and orientation is not, however, likely to be a good predic- occurs as day length begins increasing, cue for visiting bees (Jones and Buchmann tor of plant age (Grice et al. 2004, R.D. van provided conditions are warm. Pinnules 1974). Klinken unpublished data). appear to be relatively short-lived, surviv- Inflorescence maturity is acropetal, ing a few months. Facultative defoliation with pods that form at the base of the in- Perennation of rachises occurs in cooler regions such fructescence maturing fi rst (Wyatt 1981). No direct measurements of age have been as central Queensland or central Australia. In Costa Rica there were on average 9.06 recorded, although P. aculeata can be rea- Leaf loss also occurs in response to other ovules per pod (range: 7–11) and 1.7 seeds sonably long-lived in the absence of stress- stresses, such as prolonged inundation, per pod (range: 1–9) (Wyatt 1981). In Aus- es (probably >30 years). Nevertheless, fi eld drought, or insect attack. tralia P. aculeata pods average 1.6 mature demographic and growth data collected Peak fl ower production generally coin- seeds per pod (range: 1–7) with little or no throughout Australia (2000–2006, R.D. van cides with the major fl ush of foliage growth. regional variation (van Klinken and Flack Klinken unpublished data) and anecdotal Flowers are also produced throughout 2008). Pod maturation occurs within ap- observations from landholders and others the remainder of the year in all but the proximately three months of bud forma- suggest that plants rarely live for longer cooler regions, but these are responsible tion (van Klinken 2005). than 20 years in Australia. for only a small proportion of total annual pod production. Soil moisture availability Seed production Physiology does not appear to affect fl owering phe- Parkinsonia aculeata starts to produce seeds Parkinsonia aculeata is noted for its drought nology as fl owering is initiated during the in its second or third summer under the and salt tolerance and shows a number of dry season, and occurs simultaneously best possible conditions, but may take physiological adaptations to deal with low in adjacent wetland and uplands sites in longer (Vitelli 1995, R.D. van Klinken un- water uptake, including greatly reduced Australia (van Klinken 2005). Peak fl ower- published data). Seed production typically leaf area and green photosynthetic bark ing in Central America is from February to begins when plants are about 1.5 m tall, (Stewart et al. 1992). Xerophytic modifi ca- March, but is somewhat later in Mexico, increasing greatly with plant size at any tions of the leaf include heavy cutinization the southern USA and India (April and one site and year (J. Pichancourt and R.D. of the persistent epidermis, sunken stoma- May) (Hawkins 2001). Peak fl owering in van Klinken unpublished data). Large ta, a hypodermal layer of water-contain- Tucson, Arizona, occurred over a 3–4 week trees can produce at least 2500 seeds m−2 of ing cells and a photosynthetic cortex. Even period between April and June, with fl ow- projected canopy cover (van Klinken and though mature xeromorphic leaves may ering continuing until September (Dimmit Flack 2008, R.D. van Klinken unpublished attain osmotic values of 26.5 atmospheres, 1987). In Australia peak fl owering occurs data). Seed production varies consider- they are shed in times of drought (Woods in August-September in hotter regions ably between years, sites and tree densi- 1988). At fi rst the pinnules are shed, with (e.g. the wet-dry tropics), but is delayed ties (dense cohorts of tall narrow adults the leaf pinnae remaining to function as by a few months where winters are cool produce few seeds per unit area), and photosynthetic organs (Robertson and Lee (e.g. the arid interior) (van Klinken and trees may not fl ower at all during severe 1976); however, under stress, the pinnae Flack 2008). drought (Deveze 2004, p. 48). Variation are also dropped and photosynthesis is re- Most pods drop off the tree within six in pod production is generally strongly stricted to the green stems (Wood 1988). weeks of reaching maturity, although a positively correlated with foliage cover Bark photosynthesis has been described small proportion of pods can remain on and, therefore, tree health. However, very for the closely related P. fl orida, which is the tree for up to a year (van Klinken and stressed, almost leafl ess trees often pro- drought-deciduous and, like P. aculeata, Flack 2008). Pod fall can be slightly de- duce a fi nal crop of pods before death. has a large number of thin chlorophyl- layed for seasonally fl ooded trees due to lous branches (Adams and Strain 1969). both delayed maturation and delayed ab- Seed dispersal Products of photosynthesis of the stem scission (van Klinken 2005). Pods are in- Pods fall close to the parent plant. Pods cortex are similar to those of leaf photo- dehiscent and must decay or be physically can fl oat for up to 14 days (Stewart et al. synthesis. Leafl ess trees under drought damaged to release the seed. Under warm 1992, Hawkins et al. 1999, although no data stress had considerable photosynthetic wet conditions seed release can take a few were provided), and the main natural dis- activity, at least during the non-growing months, but a small proportion of pods persal mechanism for pods appears to be period, with photosynthesis at 86% of that can remain intact for over a year in arid water. Long distance fl otation may have on foliated (unstressed) trees growing in a environments. aided its spread in certain areas, and P. similar area in mid-winter. P. aculeata is a aculeata is thought to have established on light-demanding plant, its growth being Mycorrhizae and nodules Galapagos Islands (Porter 1983, in Wilson somewhat retarded in the shade (Hawkins Parkinsonia aculeata is now considered not and Miller 1987) by this means. Also, iso- 2001). to fi x nitrogen (Sprent and Sutherland lated populations found on sandy frontal Seedlings of P. aculeata are remarkably 1990, Stewart et al. 1992), although earlier dunes on the Gulf of Carpentaria in both tolerant of drought and soil variation. it was reported to do so (Hughes 1986). Queensland and the Northern Territory In a glasshouse experiment P. aculeata are thought to have arisen from pods that 106 Plant Protection Quarterly Vol.24(3) 2009 reached there through westward tidal south-western Australia, cannot be ex- which seeds are likely to be exposed to wet movements along the Northern Territory plained in this way. The importance of conditions in summer rainfall regions, and coastline from Queensland. Within catch- transported livestock (and vehicles/stock the temperatures at which seed germina- ments the importance of water dispersal is carriers) in dispersing seeds (most prob- tion is greatest (see later). evidenced by the way infestations typical- ably in mud) still needs to be resolved. Seeds from different Australian popu- ly follow drainage lines, and the frequent lations were tested, and the relationship emergence of seedlings along fl ood lines Physiology of seeds, germination and between wet heat and dormancy release and among fl ood debris. Even in uplands seed banks differed by up to a 6.6°C phase shift in the habitats, water runoff can be suffi cient to Dormancy and imbibition Parkinsonia Y-intercept (Figure 8b). This is likely to sig- spread pods (and seeds) short distances aculeata has a relatively simple dorman- nifi cantly affect the seed bank dynamics of from beneath the canopy of parent trees, cy mechanism (van Klinken and Flack different seed populations (van Klinken aiding thicket formation. 2005, van Klinken et al. 2006, van Klinken, and Flack 2005). The cause (and signifi - Although herbivores have been impli- Lukitsch and Cook 2008). The seed coat cance) of this phase shift is unknown, but cated in the dispersal of seeds through confers hard-seeded (physical) dormancy it may be related to seed moisture con- their dung, we have found little evidence through tightly packed palisade cells that tent. to support this claim. Pods have limited are impregnated with water-repellent sub- Intense dry heat from fi res will have an nutritional value and therefore are not stances, while the embryo remains fully important impact on seed dormancy (Scott adapted to dispersal by vertebrate her- developed and non-dormant (Baskin and 2006). It is likely that suffi cient dry heat bivores, unlike mesquite and prickly aca- Baskin 1998). No physiological seed dor- (e.g. 125°C for 8 minutes) will kill seeds cia (van Klinken and Campbell 2001). In mancy mechanisms have been observed while exposure to more intense heat for a feeding trial pigs did not eat P. aculeata (Everitt 1983, Alvarez Rangel 1984a, van brief periods (e.g. 300°C for 1 minute) will pods unless they were coated in molas- Klinken and Flack 2005). Once the seed release seeds from dormancy by altering ses (B. Lynes unpublished data, in Deveze is released from dormancy, imbibition seed coat properties. Fire, depending on 2004) and seeds and seedlings are rarely, if (where seeds swell to approximately its intensity, is therefore likely to kill sur- ever, reported from the dung of livestock. twice their original size) will occur pro- face seeds, releasing from dormancy some Nonetheless, incidental consumption of vided there is suffi cient moisture (van seeds that are close to the surface (<2 cm mature pods by livestock is possible. If Klinken and Flack 2005). Light is not re- in depth), and not affecting seeds that are pods were to be consumed a signifi cant quired for germination (Everitt 1983, van deeply buried. proportion of seeds would probably re- Klinken unpublished data), but seeds will main viable; for example, in the pig-feed- not germinate unless exposed to air. There Germination The relationship between ing trial, 50% of seeds that passed through is no evidence of secondary dormancy in germination and temperature has been de- the digestive tract were viable, after taking the event that germination conditions are scribed by van Klinken and Flack (2005) 3–8 days to pass (B. Lynes unpublished not met. Timing of dormancy release is, for imbibed seeds under saturated condi- data, in Deveze 2004). therefore, critical to ensure that germinat- tions. Non-dormant P. aculeata seeds ger- Birds have also been implicated in dis- ing seeds have the greatest probability of minated between 15 and 40°C. Germina- persal (Lukitsch and Wilson 1999). How- survival. tion was limited by cold stress at 20°C and ever, seed drop data collected from litter Most fresh, field-collected seeds are below and heat stress above 35°C, result- traps placed under adult trees suggests typically viable (about 100%), and most are ing in increased seed death and decreased that birds strip the seeds from almost, or dormant (>80% when inundated at 20°C) germination rates. Optimal conditions for newly, matured pods, subsequently dis- (Chatterji and Mohnot 1964, Everitt 1983, germination were between approximately carding the seed coat (R.D. van Klinken Mohnot and Chatterji 1965, van Klinken 25 and 35°C, with approximately 85% of unpublished data). No viable seeds would and Flack 2005, van Klinken, Lukitsch non-dormant seeds successfully complet- therefore be ingested for subsequent dis- and Cook 2008). P. aculeata seeds can be ing germination (defi ned as the radicle be- persal. released from dormancy by a wide range ing as long as the seed) in a median time of Parkinsonia aculeata does, however, have of mechanisms, including intense dry heat 2.1–2.9 days, depending on temperature. a surprisingly widespread distribution (Scott 2006), physical damage to the seed These results generally agree with those in Australia, which can not be explained coat, and artifi cial methods such as boil- from other germination trials (Chatterji solely by intentional plantings and dis- ing and acid scarifi cation (Mohnot and and Mohnot 1964, Alvarez Rangel 1984b, persed by water. For example, infestations Chatterji 1965, Egley 1989, Teketay 1996). Teketay 1996, Cochard 1999). often occur scattered across upland habi- Under fi eld conditions, wet heat (exposure Seeds can germinate at any time of the tats where no plantings have occurred. to wet, warm to hot conditions) has been year in Australia provided suffi cient mois- The most likely explanation is that seeds shown to be the primary mechanism (van ture availability and suitable temperatures; are dispersed in mud sticking to animals Klinken et al. 2006, van Klinken, Lukitsch however, most germinate in the summer and machinery. The use of P. aculeata trees and Flack 2008). Most dormancy release, during periods that are suffi ciently hot by livestock and other animals for shade, and subsequent imbibition, occurs within and wet to result in dormancy release and and the typically high seed densities un- one day of inundation. Very few addi- subsequent germination. In inundated der parent trees, support this hypothesis. tional seeds will imbibe after four days habitats germination occurs either in the Similarly, scotch broom (Cytisus scoparius of inundation at a particular temperature early wet season (when most or all seed- (L.) Link) seeds, which are similar to P. ac- (Figure 8a). There is also a strong, positive, lings are likely to subsequently drown) or uleata, have been shown to be dispersed relationship between temperature and as waters recedes. in mud on pig’s feet (G. Watson personal dormancy release when seeds are inun- communication 2005). dated (Figure 8a). This relationship can be Seed bank longevity Seeds will remain At a wider scale, historical, intentional described by a log-linear function above a viable and dormant indefi nitely if stored dispersal of seed by humans is likely to threshold temperature of approximately in cool, dry conditions (van Klinken, be the most important means of long- 33°C and a quadratic function below that Lukitsch and Cook 2008, R.D. van Klinken distance dispersal across Australia and threshold (van Klinken and Flack 2005). unpublished data). In contrast, most seeds between countries. However, some outly- Dormancy release is most sensitive to tem- in the seed bank will lose dormancy and ing infestations, such as a single isolated peratures between approximately 25 and either germinate or die within one year plant found recently on the roadside in 35°C, which are also the temperatures at under fi eld conditions, although rates and Plant Protection Quarterly Vol.24(3) 2009 107 (a) wet season at upland sites in the Victoria 1 River District, with only 3% of seeds re- Day 1 maining ungerminated after 35 days. The Day 4 longest-lived seeds were those in wetlands 0.8 (9% of seeds still dormant after almost Day 14 four years) and those lying fully-exposed on the soil surface (57% of seeds after two 0.6 years). The combined results from this research demonstrate that the sensitivity of the seed coat to wet heat provides a mechanism to 0.4 ensure that dormancy release will occur rapidly when conditions are optimal for seed germination and subsequent recruit- 0.2 Proportion of seeds imbibing imbibing seeds of Proportion ment (i.e. warm to hot, wet conditions, when seeds are not inundated, when seeds are buried up to about 3 cm in the soil, and 0 when there is little or no competition from 10 20 30 40 50 60 the herbaceous or canopy layer). Con- versely, seeds will be longest-lived where Temperature (°C) conditions are most buffered from high temperatures (e.g. under parent trees) or (b) are relatively dry (e.g. on the soil surface). 1 These environmental seed ‘refuges’ can Victoria River District become a source of future recruitment if Barkly Tablelands conditions change, for example following 0.8 the death of canopy trees or removal of the Central Queensland herbaceous layer.

Seedling establishment 0.6 Parkinsonia aculeata seedlings emerge as a single thorny stem. Seedlings can grow very rapidly, reaching an average height of 0.4 1.3 m in 18 months in an area of Pakistan with a mean summer temp of 32.5°C and an average rainfall of 180–200 mm (Sheikh et al. 1985), or 2 m or more per year ‘in 0.2 Proportion of seeds imbibing youth’ with ample water (Dimmit 1987); however, such growth rates are rarely ob- served in Australia (R.D. van Klinken un- 0 published data). 10 20 30 40 50 60 Recruitment rates for P. aculeata are typ- ically very low. Estimates based on seed Temperature (°C) rain, seed banks and annual demography surveys conducted at populations across Figure 8. The relationship between water temperature and the proportion of Australia (2000–2007) average 0.1% (J. seeds that are released from dormancy and subsequently imbibe: (a) with Pichancourt and R.D. van Klinken unpub- increasing inundation time and (b) for three Australian seed populations lished data). This includes some popula- after four days of inundation (after van Klinken and Flack 2005). tions that had few or no recruits over a period of several years, despite substantial seed inputs. Factors resulting in high seed- ling mortality include desiccation through lack of follow up rains, inundation, cold timing of dormancy release will depend between treatments (Figure 9a). Over 80% frosts, fi re, competition and failure of the on the microclimate the seed is exposed of seeds were still dormant at 3 and 20 cm shoot to reach the soil surface (e.g. through to, specifi cally wet heat (van Klinken et depth under dense grass and under cano- deep burial) (R.D. van Klinken unpub- al. 2006, van Klinken, Lukitsch and Cook py cover. In contrast, in the open treatment lished data) and possibly pathogens (Toh 2008). There may also be population-level (no grass or canopy cover) less than 10% of et al. 2008). differences in the responsiveness of seeds seeds were still dormant after 14 weeks at Competition with grass results in deli- to wet heat (Figure 8b). 0 cm (when covered by a thin fi lm of soil) cate thin-stemmed, slow growing P. aculea- A seed burial trial conducted over a 14 and at 3 cm depth. ta seedlings with reduced stress tolerance week period during the 2002–3 wet season A national seed burial trial was con- (J.R. McKenzie unpublished data). There is in Darwin (wet-dry tropics) (van Klinken ducted between 2001 and 2005 to deter- indirect evidence that agile wallabies (Mac- et al. 2006) compared a wide range of habi- mine seed bank longevity in different cli- ropus agilis (Gould)) selectively browse P. tats (from open to dense grass cover and mate zones (wet-dry tropics to arid), habi- aculeata from soon after emergence until forest cover) and soil depths (from 0 to tats (soil types and inundation levels) and plants are too tall to reach (about 2 m). This 20 cm) (Figure 9a). Micro-climate, through burial depths (van Klinken, Lukitsch and is likely to be a primary reason for the lack its effect on wet heat, explained the up Cook 2008) (Figure 9b). Dormancy release of recruitment at one heavily-browsed re- to 16 fold difference in dormancy levels was quickest for seeds buried during the search site in the Victoria River District 108 Plant Protection Quarterly Vol.24(3) 2009 (a) High recruitment events can occur, but these are generally rare and give rise to 100 Dormant after 2 wks Dormant after 6 wks small dense patches (10s of square metres). Dormant after 14 wks They generally occur where seed banks are high, moisture conditions are ideal, 80 and when there is limited herbaceous cover through the wet season. Large scale recruitment has been observed in exten- 60 sive areas following fl ooding, such as on fl oodplains, where in some instances P. aculeata was not previously present and 40 seeds had been transported and deposit- Seed dormancy (%) ed from infestations located higher in the catchment. 20 Vegetative reproduction Parkinsonia aculeata does not reproduce 0 vegetatively; however, it can reshoot 0 cm 3 cm 20 cm 0 cm 3 cm 20 cm 0 cm 3 cm 20 cm from basal buds located up to 20 cm be- Open Ground Canopy low ground level following damage to the cover cover main stem. P. aculeata can also be propa- gated by root or shoot cuttings (Singh (b) 1989, in Hawkins 2001). 15 Hybrids After 1 year Parkinsonia aculeata is the only Parkinsonia After 2 years species known to have naturalized in Aus- tralia, and no hybrids have been recorded 10 there; however, P. aculeata hybridizes read- ily with other Parkinsonia species in the Americas, including P. microphylla and P. praecox (Hawkins et al. 1999). P. aculeata × P. microphylla hybrids are fairly common wherever the two co-occur on the outskirts 5 of Tucson and other desert cities (Dimmit Proportion dormant (%) 1987). These hybrids grow nearly as fast as P. aculeata, are fertile, and have extremely variable progeny (Dimmit 1987). P. aculea- ta × P. praecox hybrids are common within 0 the area of sympatry of P. aculeata, and the geographically more narrowly distributed Top End (NT) VRD (NT) Barkly (NT): Barkly (NT): Pilbara (WA) P. praecox (Hawkins 1997, Hawkins et al. black soil red soil 1999). The hybrid is fertile, and is prob- ably the result of the increasing range of Figure 9. Seed bank longevity in upland habitats a) in a range of habitats P. aculeata (Hawkins 1997). This hybrid and soil depths in the wet-dry tropics (near Darwin) (after van Klinken et has been formally named Parkinsonia × al. 2006) and b) in different climatic regions and soil types when buried at carterae. P. aculeata × P. praecox is some- 2 cm (after van Klinken, Lukitsch and Cook 2008). times favoured over either parent species as a fodder, shade and ornamental tree, and is thornless, unlike either parent tree (Hawkins 1997). The ornamental trade has that was fenced to exclude cattle (R.D. van Parkinsonia aculeata seedlings of all intentionally created hybrids, including Klinken unpublished data). Similar obser- ages can survive considerable damage. ‘Desert Museum’ which is probably a (P. vations have been made for goats (East If stripped of leaves they resprout from aculeata × C. microphylla) × C. fl orida hybrid Timor, R.D. van Klinken unpublished ob- the leaf axils; if cut off above the cotyle- (Dimmit 1987). servation) and camels (Queensland, De- donary node they resprout at this node veze 2004, p. 36). Cattle also may impact (usually two shoots, one being dominant) Population on recruitment when alternative food (Woods 1988). Seedlings can even survive In Australia P. aculeata can form near- sources are limited. In central Queensland being cut off below the cotyledonary node, impenetrable thickets and thorn forests cattle have been observed to browse seed- although survival is lower and fi nal dry (Figure 10a,b); however, with relatively lings and juveniles during drought or late weight and length are signifi cantly less few exceptions (e.g. within ephemeral in the dry season, often killing the plants than when seedlings are cut off above the wetlands of the Queensland Gulf), high (Deveze 2004). In contrast, seedlings in cotyledons (Woods 1988). Most seedlings density patches are fairly small and local- the Barkly Tablelands showed little or no do, in fact, suffer tip damage, as is evi- ized within the landscape. More often, P. signs of being grazed, even when grow- denced by high rates of multi-stemming aculeata exists as sparse populations of ing in holding yards (R.D. van Klinken among seedlings less than one year old mainly mature plants (Figure 10c). Dis- unpublished data). Unidentifi ed locusts (R.D. van Klinken unpublished data). This persal, seed bank dynamics, recruitment have been observed to strip bark from slows growth rates, and must affect the success and adult longevity are all likely seedlings, sometimes causing death. competitiveness of seedlings. to be important in determining where and Plant Protection Quarterly Vol.24(3) 2009 109 when thicket formation occurs, and for Detrimental how long such thickets persist. Landscape (a) Most of the detrimental effects of P. ac- models are currently being developed us- uleata stem from its propensity to form ing Bayesian Belief Networks to predict in dense, thorny, impenetrable thickets along which parts of the landscape P. aculeata has drainage lines, depressions, ephemeral the potential to develop dense stands (van wetlands and, to a lesser extent, uplands Klinken, Seabrook et al. 2008). across a large part of Australia (Figure 5). Seed dispersal is critical in determining These are of both environmental and eco- the way P. aculeata populations are distrib- nomic signifi cance. uted and structured across the landscape. The greatest environmental impact Where the landscape is not inundated by is probably through the exclusion of the water the seed bank is largely restricted (b) herbaceous layer (van Klinken 2006). P. to under, or just adjacent to, parent trees. aculeata trees are relatively shallow-root- This provides an opportunity for slow ed, but they may shorten the duration thicket formation, especially following the that ephemeral water bodies hold water. death of parent trees. The presence of iso- Dense patches are rarely greater than 1 ha lated trees across upland habitats almost so impacts on biodiversity are likely to be certainly refl ects low rates of long distance localized and limited to the infestation site dispersal, such as in mud carried from un- (van Klinken 2006). At greatest risk are cli- der parent trees on the hooves of animals. matically suitable mesic habitats in arid In contrast, fl ooding results in the redis- and semi-arid regions, such as wetlands tribution of seeds from under parent trees, on the Barkly Tablelands (Northern Terri- provided it occurs at or soon after pod fall. tory), wetlands and gorges in the Pilbara Redistribution can occur in a number of Region (Western Australia) (van Klinken ways depending on hydrology and other (c) 2006) and waterbird habitats of national factors. For example, seeds can often be signifi cance across its potential distribu- distributed across a fl ood plain, resulting tion (Humphries et al. 1991). P. aculeata has in recruitment of relatively evenly-spaced become a problem in the wetlands of Palo new plants. Alternatively, seeds may be Verde National Park in Costa Rica where it concentrated by eddies or micro-topogra- is native, with parts of the Palo Verde La- phy into restricted parts of the landscape goon becoming extensively invaded since (Hawkins et al. 1999), resulting in the for- the early 1980s (Trama 2004). mation of discrete thickets. Finally, seeds In production systems P. aculeata can can be washed down watercourses, re- also replace pasture, but existing infesta- sulting in the isolated recruitment of new tions probably do not occur at a suffi cient plants. The latter is common where fl ood Figure 10. Invasive Parkinsonia scale to cause signifi cant and widespread waters fl ow quickly, such as within the aculeata populations: (a) aerial reductions in carrying capacities (van banks of major rivers. view of a dense infestation in Klinken 2006). Thicket formation does, Patches of P. aculeata, whether seedlings an ephemeral wetland in the however, interfere with stock manage- or adults, frequently appear to consist pri- southern Gulf of Carpentaria region ment, impedes stock access to water, makes marily of plants of a similar age. This is the maintenance of water points diffi cult (Queensland), (b) understorey of a likely to be the consequence of both large and provides refuge for feral pigs (Deveze seed banks (such as under parent trees) dense infestation in the Kimberley 2004). Both the formation and control of and the infrequency of favourable recruit- (Western Australia), and (c) a sparse thickets may also exacerbate erosion prob- ment conditions. Suitable micro-sites can population of adults in the Top End lems (Wilson and Miller 1987). Thorns may often be provided through the death of (Northern Territory). injure hooves of animals and affect leisure parent trees or receding fl ood waters. and recreational activities, while its fl ow- Parkinsonia aculeata populations can ers are known to cause hay fever (Wilson be very dynamic. For example, anecdotal and Miller 1987, Deveze 2004). In northern evidence suggests that thicket formation Mexico infestations choke irrigation chan- and decline is cyclical in parts of central severe frost (central Australia), as well as nels (Hawkins 2001). Queensland. A site in Nicaragua was ob- unexplained causes (most sites). Accumu- served to decline between 1999 and 2004 lating evidence suggests that P. aculeata Actual and potential costs to Australia when all trees were old, many were dead dieback, caused by one or more soil-borne Although P. aculeata is already widespread or dying and no young trees were present pathogens (Diplock et al. 2006, 2008) may in Australia, existing infestations are not (T.A. Heard unpublished data). Similarly, be an important driver of the often dy- yet of suffi cient scale to cause substantial of over 23 demographic sites set up in namic nature of P. aculeata patches in Aus- production losses at the property scale or healthy P. aculeata populations across Aus- tralia. A possible mechanism is that the to cause catchment or regional scale envi- tralia in 1999–2001, most had declined in dieback pathogens build up within new P. ronmental impacts. Most of the direct costs adult density over the following 4–5 years aculeata populations until they reach a con- are related to increased property manage- (J. Pichancourt and R.D. van Klinken un- centration where they are lethal to existing ment costs, especially in relation to muster- published data). This was a consequence plants and to new recruits. ing, accessing water points and maintain- of both high mortality of existing plants ing vehicle tyres, and on-ground control and low (or zero) recruitment rates, de- Importance work to prevent P. aculeata from becoming spite high seed inputs. Mortality factors Parkinsonia aculeata is an example of a a more serious problem. Costs to Austral- for adults and juveniles included dieback plant that is both weedy and benefi cial; ia will increase dramatically if P. aculeata (Barkly Tableland), fi re (Kimberley), ter- however, in Australia its negative aspects continues to spread and thicket formation mites (Kimberley, Top End), wet stress far outweigh any actual or potential ben- continues. However, actual and potential (Victoria River District wetland sites), and efi ts. impacts have not been quantifi ed. 110 Plant Protection Quarterly Vol.24(3) 2009 Benefi cial Legislation (NWS 2000). Considerable federal funding Parkinsonia aculeata is widely used as an Parkinsonia aculeata has been declared in was made available to help address the on- ornamental in dry areas throughout the all states and territories. In Queensland ground management, research and exten- Americas because of its spectacular bright it is classifi ed as a Class 2 declared pest sion priorities identifi ed in this strategy. yellow fl owers; however, it is not generally (landholders must take reasonable steps For example, approximately $A2 million considered to produce particularly valu- to keep land free of the weed; it is also of federal funding was directed towards able or high quality products (Hawkins prohibited to introduce, feed, keep, re- management of P. aculeata between 2001 2001). Uses include hedges, windbreaks, lease, take for commercial use, supply or and 2005. This was matched by at least the shade, fuel (fi rewood and charcoal), pa- transport). In the Northern Territory the same amount of cash or in-kind contribu- per-making and low quality fodder (Anon species is classifi ed as Category B (growth tions from other stakeholders (Martin and 1972, Allen and Allen 1981, MacDicken and spread to be controlled). In Western van Klinken 2006). This funding provided and Brewbaker 1984, Hughes 1989, Bar- Australia it is declared as P1 (prevention a catalyst for major on-ground control bosa and Prado 1991, Stewart et al. 1992, of trade, sale or movement), P2 (eradicate) programs in Queensland, the Northern Hawkins 2001). Although wood can be or P4 (contain) according to districts. In Territory and western NSW. In the Desert used for carpentry (Barbosa and Prado New South Wales it is declared in Cate- Channels Region (in Queensland) alone, 1991), it is brittle and of dubious durabil- gory W1 (presence must be notifi ed to the approximately 300 000 ha of P. aculeata ity (Stewart et al. 1992). P. aculeata has been local control authority and the weed must infestations as well as occurrences along used in folk medicine (Barbosa and Prado be fully and continuously suppressed and 500 km of drainage lines were treated in 1991). Leaves, when made into an infu- destroyed). In South Australia P. aculeata is 2002–2003. sion, are considered in some areas to have notifi able throughout the state, and plants A range of control techniques are now medicinal and antiseptic properties and must be destroyed. It is not naturalized available to manage P. aculeata, including the infusion has been used to treat fevers, in Victoria, the ACT or Tasmania, where the use of herbicides, machinery, fi re and epilepsy and vomiting (Stewart et al. 1992, it is declared as a restricted or prohibited grazing (Vitelli 1995, Deveze 2004). Choice Hawkins 2001). Raw seeds have been used plant. of the best or most appropriate method as a food source by humans in Mexico, depends on a number of factors, including children have been reported to eat fl owers Weed management seasonal conditions, the type of country and seeds in West Africa, and seeds have History infested, extent of the infestation, identi- been investigated as a minor food source Parkinsonia aculeata was already recog- fi ed environmental or economic risks and in India (Hawkins 2001). nized as a threat in Australia at the start of the resources available (Deveze 2004). Ir- The fodder value of P. aculeata pods and the twentieth century (Bailey 1906). Con- respective of what control method is used, foliage varies, and reports range from it trol work was commenced at least by 1940 high densities of new recruits can appear being rarely eaten by livestock or wildlife in the Fitzroy Shire (Queensland) (N. Mills following initial control, and it will be (Everitt 1983) to being a potentially im- unpublished report March 2000). Research necessary to undertake several follow-up portant fodder tree (MacDicken and Brew- into P. aculeata control began as early as treatments. backer 1984, Stewart et al. 1992, Hawkins the 1950s, and the fi rst recommendation 2001). It appears to be consumed by cattle was 2,4,5-T in diesel (NWS 2000). This was Herbicides only in times of shortage (Stewart et al. the standard method of control in all areas Presently, there are a limited number of 1992), such as late in the dry season (Anon. of infestation on properties and govern- herbicides available for the control of P. 1972, Deveze 2004, p. 35, 45); however, it is ment reserves until the chemical stopped aculeata in Australia (Table 1) (Vitelli 1995, browsed by sheep, goats and camels and, being used in the early 1990s. In addition Deveze 2004); however, those that are reg- in some parts of the world, branches are to this, some landholders utilized fi re and istered are highly effective when applied lopped during dry periods to feed sheep mechanical control. A biological control at the recommended life stages and under and goats (Hawkins 2001). program, jointly funded by Queensland, favourable climatic conditions (Deveze Parkinsonia aculeata has been introduced the Northern Territory and Western Aus- 2004, McKenzie et al. 2004). Selection of the pantropically, primarily as an ornamen- tralia, commenced in the 1980s (Woods most appropriate method of application tal, hedging and fodder tree (Stewart et al. 1986). Research on chemical control meth- is very important, with options available 1992, Woods 1988, Hawkins 2001). In ad- ods also recommenced in the early 1980s including basal bark and cut stump spray- dition, its tolerance to drought, waterlog- as the weed problem became more seri- ing, and ground and aerial applications ging and saline conditions has meant that ous. Major on-ground control work also of both foliar and soil applied herbicides it has often been promoted for rehabilita- commenced around that time and has con- (Table 1) (Vitelli 1995, Deveze 2004). tion and as a multi-purpose tree, particu- tinued through to today. For example, in Basal bark spraying and cut stump larly in harsh, degraded or marginal land Queensland $A233 706 was spent on eight applications are best suited for treating (Hughes 1986, Hawkins 2001). It has been Strategic Weed Eradication and Education scattered to light infestations. Because of used for reforestation programs in several Program projects in fi ve shires between the high labour input required with these countries, including India, Sudan and 1995 and 1998 (NWS 2000). techniques, treatment of denser infesta- Cape Verde (Le Heoerou 1984, Hughes In 1999 P. aculeata was recognized as tions becomes less economical. For ex- 1989) and continues to attract attention as one of the 20 Weeds of National Signifi - ample, in a large integrated trial under- a candidate for the reforestation of degrad- cance, i.e. as one of the worst weeds in taken on P. aculeata, basal bark spraying ed environments (Madany 1991, Rajaram Australia (Thorpe and Lynch 2000). As a of a dense infestation (average density of and Janardhanan 1991, Stewart et al. 1992). consequence, in 2000 a National Strategic 2200 plants ha−1) cost $A420 ha−1 (McKen- However, its usefulness can be limited by Plan was developed which aimed to de- zie et al. 2004). Best results for basal bark its weedy tendencies (Hughes 1989). liver four key outcomes: (i) P. aculeata man- treatment are achieved when the plants In Australia P. aculeata appears to have agement is coordinated and maintained are growing actively, whereas cut stump been planted mainly as an ornamental and at a national level; (ii) Zone A infestations treatment is effective throughout the year shade tree. Its potential for use in rehabili- (containment zones) are reduced; (iii) (Deveze 2004). tation of Australia’s arid regions has been Zone B infestations (active control zone) Foliar spraying of the foliage and green discussed (Ratcliffe 1936, in Cook and Dias are minimized; and (iv) Zone C (eradica- stems of P. aculeata is a highly effective 2006), but we are not aware of it ever being tion zone) infestations are eradicated and control method for treatment of actively used for that purpose. P. aculeata is prevented from spreading growing seedlings and young plants up Plant Protection Quarterly Vol.24(3) 2009 111 Table 1. Registered herbicides for Parkinsonia aculeata control in AustraliaA (adapted from Deveze 2004 and Infopest July 2008, Biosecurity Queensland, Department of Primary Industries and Fisheries, Brisbane). Active chemical constituents Application method States/Territories where Rate registered Tebuthiuron (200 g kg−1) Aerial Qld and NT 10–15 kg ha−1 Soil Qld, NT and WA 10–15 kg ha−1 (1–1.5 g m−2)

Triclopyr (300 g L−1) + picloram Aerial (helicopter only) Qld and NT 3 L ha−1 (100 g L−1) Foliar (overall spray) Qld, NT and WA 350 mL 100 L−1 water

Triclopyr (300 g L−1) + picloram Aerial (helicopter only) Qld, NT and WA 3 L ha−1 (100 g L−1) + aminopyralid (8 g L−1) Foliar (overall spray) Qld, NT and WA 350 mL 100 L−1 water

Triclopyr (240 g L−1) + picloram Basal bark and cut stump All states/territories 1 L 60 L−1 diesel (120 g L−1) Hexazinone (250 g L−1) Soil – spot spray (individual tree) Qld, NSW, WA and ACTB 4 mL per spot, 1 spot for each bush or tree Soil – spot spray (grid pattern) Qld, NSW, WA and ACTB 1 mL per spot to 1m tall, 4 mL per spot to 5 m tall A Refer to individual product labels for details on the situations in which the products can be sprayed and the best times to spray. B Velpar® L is not registered for use in the ACT.

to a height of 1.5–2 m (Deveze 2004). As applying a liquid herbicide ($A154 ha−1) enable plants to be dug out with a large for basal bark and cut stump applications, (McKenzie et al. 2004). The ground appli- portion of root. This would, however, take foliar spraying becomes expensive as the cation of hexazinone proved much more longer, thereby increasing the cost of treat- density of P. aculeata increases. For dense expensive ($A270 ha−1) and is more appro- ment. Use of tractors mounted with blades infestations, aerial application of foliar priate for scattered to light infestations. or small cutter bars could be a feasible op- herbicides can be undertaken, but effi cacy tion for treating scattered infestations of may vary depending on the density and Mechanical options plants, as has been used for treatment of size of the plants. McKenzie et al. (2004) Parkinsonia aculeata can be killed using some prickly acacia infestations in west- reported only 66% mortality of a dense mechanical techniques, some of which be- ern Queensland. In central Queensland infestation following aerial spraying and come more economical as the density of a small Ellrott plough has been attached concluded that this may have been be- infestations increases. Available options to a Bobcat and is proving very effective cause there were many plants larger than 2 include grubbing, stick raking, cutter bar- at treating isolated P. aculeata plants and m in height located within the fi eld site. In ring, blade ploughing and chain pulling other woody weeds. a trial undertaken to quantify the effi cacy (Deveze 2004). Costs of control have been The cheapest mechanical option to im- of aerial spraying, it was concluded that estimated in recent trials (see below), al- plement for dense infestations of woody there was a direct inverse relationship be- though it should be noted that diesel prices weeds is chain pulling. However, P. aculea- tween plant size and mortality (J.S. Vitelli have increased considerably since then. ta appears to be quite tolerant of this tech- unpublished 2006). As for many woody weeds, mechanical nique, with results from one study fi nd- The use of residual herbicides is fa- clearing of P. aculeata is most successful ing that less than 30% mortality occurred voured by many land managers who when the roots of plants are removed or even after infestations were chained suc- prefer not to revisit treated areas and con- sheared off below the bud zone (20–30 cm cessively in opposite directions (double tinually treat new recruits (Landsberg et below ground) (Deveze 2004). McKenzie pulled) (McKenzie et al. 2004). The main al. 2006). Such herbicides can be applied et al. (2004) found that blade ploughing benefi t of chain pulling is that plants may through ground based applications, or was extremely effective for treating dense be prevented from seeding for a period of aerially. For the two registered products, infestations, with greater than 90% mor- time, thereby reducing replenishment of hexazinone and tebuthiuron, McKenzie et tality achieved, irrespective of whether the seed bank. al. (2004) achieved 97% and 81% mortality a back mounted or front mounted blade In many instances, broad scale mechan- respectively, in a dense infestation after plough was used. The front mounted ma- ical control of woody weeds will cause one season. After two wet seasons mor- chine was slightly cheaper at controlling considerable disturbance and potentially tality further increased to 99% and 95% P. aculeata ($A126 ha−1) on a per unit area create an environment conducive to large respectively. Hexazinone was applied in- basis compared with the back mounted scale seedling recruitment. In the absence dividually to each plant for medium to option ($A156 ha−1), probably because of of any follow up activities the problem can scattered infestations and as a meter-by- its greater manoeuvrability. Use of a bull- soon become worse than it initially was. meter grid format for dense overlapping dozer with a standard blade achieved only Conversely, promoting a favourable envi- canopy infestations, whereas tebuthiuron 43% mortality yet cost almost as much as ronment for germination can help deplete was broadcast from an aeroplane. The blade ploughing. McKenzie et al. (2004) the seed bank more quickly than normal cost associated with the broadscale ap- suggested that if bulldozers with standard and shorten the length of time during plication of tebuthiuron averaged $A180 blades are all that is available, careful ma- which follow up control may be needed ha−1, slightly more than that for aerially noeuvring of the blades at an angle may (Campbell and Grice 2000). 112 Plant Protection Quarterly Vol.24(3) 2009 There are some legislative restrictions of 820°C killed 83% of the treated plants If mature pods are present on plants, that may infl uence whether mechanical at a cost of 7.5 cents per plant (Vitelli and camels may consume them inadvertently techniques can be used in certain situa- Madigan 2004). This practice would be or if there is no other preferred food avail- tions. For example, in Queensland, it is most appropriate for treating small and/ able (McKenzie et al. 2006). Camel dung generally necessary to obtain a permit un- or scattered infestations. from within infested paddocks contained der the Vegetation Management Act 1999 only small numbers of viable seeds (av- to use machinery in riparian areas (Vitelli Grazing management systems erage of less than two seeds within 30 2000, Vitelli and Pitt 2006). There is a paucity of information in the samples), with a maximum recording of literature on the effects that grazing and 11 seeds (McKenzie et al. 2006). A comple- Fire browsing have on P. aculeata dispersal, re- mentary study suggested that approxi- Anecdotal evidence has provided mixed cruitment or control. Several landholders mately 15% of consumed seeds remain reports on the effi cacy of fi re as a control suggest that cattle will browse P. aculeata viable after ingestion by camels (J. McKen- option for P. aculeata. They include sug- foliage (Deveze 2004), but this happens zie unpublished data). From these fi nd- gestions of high mortality being achieved generally only when available grass is lim- ings it was concluded that when camels through cool low intensity burns to limited ited or of low nutritive value, such as at the are confi ned to an infested area, the threat control after hot fi res (Deveze and Kend- end of the dry season prior to fi rst storms. of seed spread is inconsequential com- rick 2004, in Deveze 2004, McKenzie et al. Under these circumstances cattle can eat pared to the benefi ts of both fewer seeds 2004, in Deveze 2004). Preliminary results whole stems to 2 cm in diameter, some- being produced and a low seed viabil- from a trial undertaken to quantify the ef- times killing the plants (Deveze 2004). ity of consumed seeds; however, in areas fi cacy of different intensity fi res (headfi res Plant analysis studies have found that P. where camels are not confi ned, they may versus backfi res) imposed in various sea- aculeata has low leaf fodder value due to act as vectors for dispersal into clean areas sons, suggest that high kill rates can be its moderate to high total fi bre content (McKenzie et al. 2006). Whether repeated achieved for all size classes of P. aculeata. levels, a nitrogen content of 3.7%, and no damage by camels could eventually cause Effi cacy does, however, vary depending condensed or hydrolyzable tannins, which mortality of adult P. aculeata plants has not on the season and intensity of burning usually indicate high digestibility (Hunt- been determined. The impacts of camels (Grice et al. 2004). er and Stewart 1993). Nonetheless, some to off-target species should also be con- It is likely that a large portion of P. animals, especially macropods, goats and sidered when contemplating their use as aculeata seeds located on or close to the camels, can preferentially browse P. ac- a tool for P. aculeata management, because soil surface will be killed by fi res (Scott uleata, substantially reducing growth and under some grazing regimes camels have 2006). Buried seeds, however, will gener- survival of seedlings and juveniles. the potential to contribute to a decline ally survive. The insulating ability of soil Parkinsonia aculeata seedlings are rela- in preferred forage species (Dorges and negates the high temperatures produced tively poor competitors, and heavy pas- Heucke 2003). by fi res, with only small increases being ture cover is therefore likely to limit re- The unexplained presence of P. aculeata recorded 1 cm below ground during most cruitment through competition for light in upland situations suggests livestock grass fi res (Bebawi and Campbell 2002, and moisture. In fact, high densities of might play a role in seed dispersal, either Campbell and Setter 2002). For surviving vigorous P. aculeata seedlings have only through dung or in mud. Until the means seeds, dry heat from burning can com- been noted where there is little herbaceous of dispersal has been confi rmed, it should promise the seed coat, releasing the seed cover. It is not yet known how important be presumed that livestock can pose at from dormancy (Scott 2006). Removal of grazing management is for reducing P. ac- least a small risk of seed dispersal, and vegetation through fi re can also result in uleata recruitment relative to other factors livestock movement from infested to clean seeds being exposed to wet, hot conditions such as fl ooding (or the absence thereof). areas needs to be managed accordingly (wet heat) following rainfall, which is the Also, it is possible that spelling pastures (see van Klinken and Campbell 2001). key dormancy release mechanism for this to promote competition may produce species (van Klinken 2005). The implica- adverse results by allowing seedlings Commercial exploitation tion for management is that removal of to grow freely in situations where cattle No P. aculeata products have been identi- herbaceous cover by burning is likely to browse this weed. The effect of grazing fi ed that can be exploited on a commercial expedite the run down of the seed bank on P. aculeata recruitment is likely to vary scale in Australia. by triggering dormancy release and sub- with habitats, for example, perennial ver- sequent germination. This is a desirable sus annual grasses pastures, riparian ar- Natural enemies scenario, provided land managers are in eas that are dominated by approximately Native natural enemies in the introduced a position to treat any seedling regrowth 2 m tall cane-grass (Chionachne cyathopoda range There are relatively few native in- before it has the opportunity to reach re- (F.Muell.) Benth), and seasonal wetlands sect herbivores on P. aculeata in Australia. productive maturity. where annual grasses might begin grow- In the Northern Territory only 17 spe- Although fi re offers an effective tool, ing before water recedes. cies of insects were found feeding on the its use is frequently limited by lack of fuel Recently, a small but increasing number weed, and none was common (C. Wilson load, social or cultural constraints, the of northern Australian landholders have personal communication 2008). The most safety risk of late dry season burns, the been using camels for woody weed con- damaging was the giant termite (Mas- desire to conserve fodder, and concerns trol (Deveze 2004). Preliminary studies totermitidae: Mastotermes darwiniensis about non-target environmental effects, undertaken to quantify their impacts Froggatt) which is capable of killing ma- particularly in wetland and riparian habi- found that camels selectively browsed ture plants, although plants may survive tats (Deveze 2004). fresh growth of P. aculeata, including fl ow- and send up fresh suckers (R.D. van Flame throwers have been tested as ers, young pods and leaves (McKenzie et Klinken personal observation). Other an individual treatment for P. aculeata al. 2006). Consequently, if paddocks are species were seven species of stem suck- where chemical and mechanical controls stocked at a suffi ciently high density they ing bugs, two pod and seed feeders, two are inappropriate or ineffective (Vitelli can reduce the quantity of P. aculeata seeds species of grasshoppers and a case moth and Madigan 2004). Flaming for 10 sec- produced to negligible levels. This in turn caterpillar that chews leaves. The com- onds around the entire circumference of results in minimal replenishment of the bined impact of all of these on the plant the base of individual plants (5 cm above seed bank and less seedling recruitment was considered to be negligible; however, ground level) at a maximum temperature in future years. there are reports of insects being locally Plant Protection Quarterly Vol.24(3) 2009 113 abundant and damaging, including an Classical biological control Investiga- and Flack 2008). Several other insects unidentifi ed coccid that reaches high den- tions into the potential biological control with high potential may deserve re-eval- sities causing the gnarling of growth tips of P. aculeata began in 1983 with a joint uation, especially the tortricid, Ofatulena (Victoria River District, R.D. van Klinken project between state government agri- luminosa Heinrich, which attacks tips and unpublished data) and stem borers (Halls cultural agencies in Queensland, West- green seeds. It was originally excluded Creek [Kimberley], N. Wilson personal ern Australia and the Northern Territory because it is diffi cult to rear and test for communication 2001). These insects often (Woods 1986). The Australian Weed Com- host specifi city, but new techniques may appear to be most abundant on plants that mittee approved P. aculeata as a target be applied to assess this insect. Also, the are already stressed from other causes. for biological control in Australia in 1983 bruchid, Mimosestes amicus Horn deserves The univoltine, pantropical, seed feeding (Donnelly 2000). reappraisal. The host range of this species moth, Mesophleps palpigera (Walsingham) Surveys were conducted from April appears to be too wide, but research on the (Gelechiidae) is widespread. It oviposits 1983 until January 1987 from bases in Texas occurrence of races or cryptic species with on tree pods just prior to seed maturation and Arizona (USA) and Guerrero (Mexico) limited host ranges may be useful. and has caused up to 33% seed mortality (Woods 1986, 1988, 1992). These surveys Rhinacloa callicrates has been released in some samples, although seed predation were conducted in the states of Texas, across northern Australia in large numbers rates are generally no more than a few per- Arizona (USA), Tamaulipus, Nuevo Leon, (500 000 in Queensland alone). Although cent (van Klinken 2005, van Klinken and Sonora, Sinaloa, Baja California, Michoa- it reaches very high, damaging densities Flack 2008). can, Guerrero, Oaxaca, San Luis Potosi, in its native range, and has established in A natural dieback phenomenon has Hidalgo (Mexico) and Guanacaste (Costa central Queensland, it has never been ob- been widely observed across the Austral- Rica). An emphasis was placed on collect- served to reach damaging densities there ian distribution of P. aculeata (excluding ing in the Sonoran desert because it was (Donnelly 2000). Failure to establish in the the Pilbara Region and Central Australia) thought to be the centre of origin of the ge- Kimberley could be because temperatures for at least 50 years (Diplock et al. 2008). It nus and other areas had not yielded many and relative humidity are too high (G. causes widespread deaths of individuals insects (Woods 1988). North America (USA Donnelly personal communication 1999), and populations of plants. In some regions and northern Mexico) was surveyed thor- although this has not been tested. it has been identifi ed anecdotally as the oughly, but only limited surveying was The two seed-feeders were released factor preventing P. aculeata from becom- conducted in Central America (including between 1995 and 1999 on P. aculeata ing an important problem. Research to southern Mexico and Costa Rica). throughout much of its range in Aus- identify the causal agent is continuing, but Two mite species and 142 insect spe- tralia (Donnelly 1998, Lockett et al. 1999, likely causative agents include Lasiodiplo- cies were collected from P. aculeata dur- Lukitsch and Wilson 1999), including dia theobromae (Pat.) Griffon & Maubl., Fu- ing the surveys (54 from Arizona, 38 from the Kimberley (Western Australia). Mi- sicoccum dimidiatum (Penz.) D.F.Farr and Texas, 69 from Mexico and 16 from Costa mosestes ulkei oviposits on nearly mature Fusarium sp., all of which are generalist Rica). Species diversity at least partly re- pods either on the tree or on the ground root-borne pathogens (Diplock et al. 2006, fl ected sampling effort. Most insects were (Woods 1985). It appears to have estab- 2008). Pathogens also cause death of newly polyphagous, with many utilizing the lished at relatively few sites and, where germinating seedlings in laboratory tests, genera Prosopis, Acacia and other species measured, the seed mortality rates have although it is not yet known what impacts of Parkinsonia. In comparison with other been low (maximum of 5.3% of seeds in they might have under natural conditions. plant species, a large number (44) of phy- the Northern Territory) (Donnelly 1998, In a laboratory study, the pathogenicity tophages were polyphagous ‘Homoptera’. Lockett et al. 1999) (see also Woods 1988). of 84 pathogen isolates extracted from P. On the basis of fi eld host records only It has not been reported in the past several aculeata plants sampled across Australia eight species (6%) were considered to be years. In contrast, Penthobruchus germaini was tested against germinating P. aculeata suffi ciently host-specifi c to be considered established easily, and dispersed readily; seedlings (Toh et al. 2008). Pathogenicity for biological control in Australia (Woods including to Alcoota station in the Alice varied greatly, but fi ve isolates from the 1988). Only the polyphagous bruchids, Springs Region several hundred kilome- Northern Territory, Western Australia and Stator limbatus (Horn) and Mimosestes ami- tres from the closest known release site Queensland (Lasiodiplodia spp., Fusicoccum cus (Horn), were widespread in North and (van Klinken and Flack 2008). P. germaini sp. and an unidentifi ed isolate) caused Central America. A few other insects were passes through several generations a year, almost 100% seedling mortality. Despite widely distributed in Arizona, Texas and and oviposits primarily on pods on the extensive sampling for internal patho- Mexico. tree (Briano et al. 2002, van Klinken 2005, gens on P. aculeata in its native range no Twelve insect species (eight from the van Klinken and Flack 2008) (Figure 11). symptoms have been found. A total of 76 Sonoran Desert Region, and the rest from There was an initial anecdotal report of sites have been sampled, with 50 stems southern USA and Mexico) were identifi ed seed mortalities exceeding 99% of avail- cut and inspected for vascular staining as potential biological control agents be- able seeds in central Queensland (Don- or other signs of pathogens at each site. cause of the damage they caused (Woods nelly 1998). However, seed consumption The sites are distributed in Mexico, Costa 1988). Several of these were excluded on rates were relatively low during national Rica, Venezuela, Guatemala, Brazil, Peru, the basis of their broad host range. The surveys conducted between 2000 and 2004 Puerto Rico, Paraguay and Dominican two with greatest potential, Rhinacloa cal- (van Klinken 2005, van Klinken and Flack Republic. licrates Herring (a sap-sucking mirid) and 2008), and the agent is therefore unlikely In a survey of the natural enemies of P. Mimosestes ulkei (Horn) (a seed-feeding to be causing any population-level im- aculeata in Kenya, only nine species of in- bruchid) were released in Queensland in pacts. Studies showed that beetle popula- sects were found and all were indigenous 1993 (Julien and Griffi ths 1998) and the tions were unable to track sudden season- polyphagous species (Robertson 1988). Northern Territory in 1989 (Donnelly al fl uctuations in pod supply, resulting in Natural enemies have also been reported 2000) and 1994 (Flanagan et al. 1996) re- a lag-phase between seed availability and on P. aculeata in India, including the seed- spectively. A third insect from Argentina, beetle numbers. Also, high egg parasitism feeding moth Enarmonia malesana (Mey- the seed-feeding bruchid, Penthobruchus (10–70%) by a trichogrammatid wasp (Us- rick), and the sap-sucking mealy-bugs germaini Pic., was identifi ed from the lit- cana sp.) is likely to be a key regulating fac- Icerya aegyptiaca (Douglas) and Pseudoao- erature as a potential agent and was re- tor through its effect on egg survival, and nidia sp. (Troup and Joshi 1983). leased in Australia from 1995 (Briano et indirectly on adult densities. Even without al. 2002, van Klinken 2005, van Klinken egg parasitism, the unexplained failure of 114 Plant Protection Quarterly Vol.24(3) 2009 6–44% of unparasitized eggs to hatch, high Several insect species, however, show Lawes (CSIRO) for their unpublished unexplained larval mortality within seeds potential as biocontrol agents. The tortri- data and observations, the many partici- (62%) and slightly aggregated (negative cid (Cochylinae) moth, Rudenia near legu- pants of parkinsonia research workshop binomial) oviposition, means that average minana Busck, is one of the most common series (2000–2006) for their contributions egg densities would need to be over 8.5 and consistently-found herbivores in the towards improving our understanding of eggs per seed to achieve 80% seed mortal- native range. The larva lives in a hole that parkinsonia, Lauren Quinn for assistance ity (van Klinken and Flack 2008). it bores in the stem tip, eventually killing with the maps, Tony Grice for comments Existing agents therefore do not appear it. This species appears to be wide ranging on a draft manuscript and the Australian to be having a signifi cant impact, and new geographically and has a wide host range Federal Government for funding assist- potential agents are unlikely to be found that includes Acacia; however, molecular ance (Department of Agriculture, Forestry in the United States or northern Mexico. sequencing of the CO2 mitochondrial gene and Fisheries). However, Central and South America indicates strong genetic differentiation of have not been surveyed properly. Native- the populations from different provenanc- References range surveying was reinitiated by CSIRO es. Furthermore, preliminary host testing Adams, M.S. and Strain, B.R. (1969). Sea- Entomology in 1999 in Mexico, Guatema- of a provenance from Mexico indicates sonal photosynthetic rate in stems of la, Nicaragua, Costa Rica and Venezuela, specifi city to P. aculeata. Hence, there is a Cercidium floridum. Photosynthetica 3, where a total of 101 sites were visited at possibility of the existence of races or cryp- 55-62. least once between 1999 and 2005 (Heard tic species with limited host ranges. Rear- Allen, O.N. and Allen, E.K. (1981). The Le- 2006). ing is not diffi cult and the feeding damage guminosae. In ‘A source book of char- The discovery that the current broad to leaves and boring damage to tips are acteristics, uses and nodulation’. (Uni- distribution and deep genetic structuring impressive. versity of Wisconsin Press, Madison, of P. aculeata refl ects very old dispersal In addition to the insects, a fungal path- Wisconsin, USA). events (Hawkins et al. 2007) stimulated ogen, Septoria sp., is considered a prom- Anon. (1972). Parkinsonia, a weed to interest in survey work in South America. ising potential biocontrol agent for P. ac- watch. Turnoff 4, 18-9. Hawkins et al. (2007) focussed on North uleata. In Nicaragua it has been observed Anon. (1978). Parkinsonia. WADA Ag- America, but the Venezuelan samples and to cause rachis and branch cankering lead- riculture Protection Board Advisory the one Argentina sample differed strongly ing to signifi cant damage and die-back (H. Leafl et, 37 pp. (although populations from Peru, Bolivia, Evans personal communication 2007). Alvarez Rangel, R.A. (1984a). Caracteriza- Paraguay and Ecuador were not included Some native-range surveys have also cion de los mecanismos de control de la in the study). This genetic information, in been conducted in Argentina and Para- latencia y germinacion de las semillas combination with observations made dur- guay by the USDA (H. Cordo personal de Parkinsonia aculeata L.1. Respuestas ing recent fi eld trips in Brazil, Ecuador and communication 2007). Potential agents de las semillas a los tratamientos para Peru, indicates that P. aculeata is native to were identifi ed, including a geometrid romper la latencia y sus implicaciones these areas and arrived there millions of moth (Briano et al. 2002). Stem-boring cer- ecologicas. Revista de la Facultad de years ago, and that these populations may ambycid beetles have also been surveyed Agronomia (Maracay, Unversidad Central harbour unique and specifi c co-evolved near Buenos Aires, Argentina (di Iorio de Venezuela) 8, 5-30. natural enemies. By early 2008, approxi- 1993). Alvarez Rangel, R.A. (1984b). Caracteriza- mately 340 species had been recorded cion de los mecanismos de control de la from a total of 190 sites in eight countries, Acknowledgments latencia y germinacion de las semillas including Ecuador, Brazil and Peru, which We thank Rob Parr, Noel Wilson (Agricul- de Parkinsonia aculeata. II. Respuestas had not been searched before. Only a sub- ture Western Australia), Alan Thompson de las semillas a algunos factores ambi- set of these is likely to be herbivorous on P. (CALM WA), Jim Begley, Grant Flanagan, entales y sus implicaciones ecologicas. aculeata (Figure 12). Disappointingly, few John Gavin, Bert Lukitsch, John McMa- Revista de la Facultad de Agronomia (Ma- appear to be damaging, wide ranging and hon, John Peart, Steve Wingrave (NT racay, Unversidad Central de Venezuela) common, attributes desired in biocontrol Weeds Branch), Ben Lawson (University 8, 31-45. agents (T.A. Heard unpublished data). of Queensland), Tony Grice and Roger Bailey, F.M. (1906). ‘Weeds and suspected poison plants of Queensland’. (H. Pole and Co., Brisbane).

Figure 12. Larvae and leaf feeding damage Figure 11. An adult Penthobruchus germaini. Image provided by on a P. aculeata growing tip caused by an Biosecurity Queensland. unidentifi ed geometrid moth in Guatemala. Plant Protection Quarterly Vol.24(3) 2009 115 Barbosa, D.C.A. and Prado, M.G. (1991). Diplock, N., Galea, V. and van Klinken, Weeds, eds V.C. Moran and J.H. Hoff- Quantitative analysis of the growth of R.D. (2008). Movement of dieback front mann, pp. 441-3. (University of Cape Parkinsonia aculeata L. in a greenhouse. through a stand of parkinsonia – a time Town, Cape Town). Phyton 52, 17-26. series study. Proceedings of the 16th Grice, A.C., Campbell, S.D., McKenzie, Baskin C.C. and Baskin J.M. (1998). ‘Seeds: Australian Weeds Conference, eds R.D. J.R., Whiteman, L.V. and Lukitsch, B.V. ecology, biogeography and evolution of van Klinken, V.A. Osten, F.D. Panetta (2002). Size-biomass relationships for dormancy and germination’. (Academ- and J.C. Scanlan, p. 186. (Queensland Australian populations of the invasive ic Press, New York). Weeds Society, Brisbane). rangeland shrub Parkinsonia aculeata L. Bebawi, F.F. and Campbell, S.D. (2002). Ef- Diplock, N., Galea, V., van Klinken, R.D. Rangelands Journal 24, 207-18. fects of fi re on germination and viability and Wearing, A. (2006). A preliminary Grice, A.C., Campbell, S.D., McKenzie, of bellyache bush (Jatropha gossypifolia) investigation of dieback on Parkinsonia J.R., Whiteman, L.V., Pattison, M., seeds. Australian Journal of Experimental aculeata. Proceedings of the 15th Aus- Headricks, K. and Andrew, M.H. (2004). Agriculture 42, 1063-9. tralian Weeds Conference, eds C. Pres- Problems of population structure exem- Benson, L. and Darrow, R.A. (1981). ‘Trees ton, J.H. Watts and N.D. Crossman, pp. plifi ed by the invasive tropical shrub and shrubs of the south western desert’. 585-7. (Weed Management Society of Parkinsonia aculeata L. in northern Aus- (University of Arizona Press, Tucson). South Australia, Adelaide). tralia. Rangelands Journal 26, 237-48. Briano, J.A., Cordo, H.A. and DeLoach, C.J. Donnelly, G.P. (1998). Levels of attack and Haston, E.M., Lewis, G.P. and Hawkins (2002). Biology and fi eld observations destruction of Parkinsonia aculeata seeds J.A. 2005. A phylogenetic reappraisal of of Penthobruchus germaini (Coleoptera: by bruchid biological control agents. the Peltophorum group (Caesalpinieae: Bruchidae), a biological control agent Proceedings of the Sixth Australasian Leguminosae) based on the chloroplast for Parkinsonia aculeata (Leguminosae: Applied Entomological Research Con- trnL-F, rbcL and rps16 sequence data. Caesalpinioideae). Biological Control 24, ference Volume 2, eds M.P. Zalucki, American Journal of Botany 2, 1359-71. 292-9. R.A.I. Drew and G.G. White, p. 310. Hawkins, J.A. (1996). Systematics of Par- Campbell, S.D. and Grice, A.C. (2000). (University of Queensland Press, Bris- kinsonia L. and Cercidium Tul. (Legumi- Weed biology: a foundation for weed bane) nosae: Caesalpinioideae.). PhD thesis, management. Tropical Grasslands Journal Donnelly, G.P. (2000). Biology and host Linacre College, Oxford. 34, 271-9. specifi city of Rhinocloa callicrates Her- Hawkins, J.A. (1997). Systematics of Par- Campbell, S.D. and Setter, C.L. (2002). ring (Hemiptera: Miridae) and its in- kinsonia and Cercidium. Oxford Forestry Mortality of Prosopis pallida (mesquite) troduction and establishment as a Institute Research Highlights 4, 1-2. following burning. Australian Journal of biological control agent of Parkinsonia Hawkins, J.A. (2001). Parkinsonia aculeata Experimental Agriculture 42, 581-6. aculeata L. (Caesalpiniaceae) in Austral- (Mexican palo-verde). In ‘Forestry com- Carter, A.N. (1974). The genus Cercidium ia. Australian Journal of Entomology 39, pendium’. (CABI Publishing, Oxon, (Leguminosae, Caesalpinioideae) in 89-94. UK). the Sonoran Desert of Mexico. Proceed- Dorges, B. and Heucke, J. (2003). Dem- Hawkins, J.A., Boutaoui, N., Cheung, K.Y., ings of the California Academy of Sciences, onstration of ecologically sustainable van Klinken, R.D. and Hughes, C.E. Fourth Series 40, 17-57. management of camels on aboriginal (2007). Intercontinental dispersal prior Chatterji, U.N. and Mohnot, K. (1964). Eco- and pastoral land. National Heritage to human translocations revealed in a physiological studies on the germina- Trust Final Report on project number: cryptogenic invasive plant. New Phy- tion of seeds of certain arid zone plants: 200046. www.camelsaust.com.au/NHT tologist 175, 575-87. Part 1 – Germination experiments with report2003.doc. Hawkins, J.A. and Harris, S.A. (1998). the seeds of Parkinsonia aculeata. Pro- Doyle, J.J., Chappill, J.A., Bailey, C.D. and RAPD characterisation of two neotrop- ceeding on Problems of Indian Arid Kajita, T. (2000). Towards a comprehen- ical hybrid legumes. Plant Systematics Zone pp. 135-9. (Ministry of Education, sive phylogeny of legumes: analyses of and Evolution 213, 43-55. Government of India, New Delhi). rbcL sequences and a general data set. In Hawkins, J.A., White, L., Hughes, C.E., Cochard, R. (1999). A seed ecology study of ‘Advances in legume systematics’, Part Contreras-Jiménez, J.L. and Mercado, Parkinsonia aculeata L. Honours thesis, 9, eds P.S. Herendeen and A. Bruneau. P. (1999). Investigation and documen- School of Tropical Environment Studies (Royal Botanic Gardens, Kew). tation of hybridisation between Par- and Geography, James Cook University Egley, G.H. (1989). Water-impermeable kinsonia aculeata and Cercidium prae- of North Queensland. seed coverings as barriers to germina- cox (Leguminosae: Caesalpinioideae). Cook, G.D. and Dias, L. (2006). It was no tion. In ‘Recent advances in the devel- Plant Systematics and Evolution 216, accident: deliberate plant introductions opment and germination of seeds’, ed. 49-68. by Australian government agencies R.B. Taylorson, pp. 207-23. (Plenum Heard, T.A. (2006). Parkinsonia aculeata: sur- during the 20th century. Australian Jour- Press, New York). veys for natural enemies, native range nal of Botany 54, 601-25 Eisikowitch, D. and Masad, Y. (1982). Pref- ecological studies and prospects for bi- Deveze, M. (ed). (2004). ‘Parkinsonia na- erences of honeybees for different or- ological control. Proceedings of the 15th tional case study manual’. (QNRME, namental nectar-yielding plants during Australian Weeds Conference, eds C. Brisbane). the dearth period in Israel. Bee World Preston, J.H. Watts and N.D. Crossman, di Iorio, O.R. (1993). Cerambycidae (Co- 63, 77-82. pp. 581-4. (Weed Management Society leoptera) y otros insectos asociados a Everitt, J.H. (1983). Seed germination char- of South Australia, Adelaide). Parkinsonia aculeata (Leguminosae) en acteristics of two woody legumes (Re- Herendeen, P.S., Bruneau, A. and Lewis, Argentina. (Cerambycidae (Coleop- tama and Twisted Acacia) from South G.P. (2003). Phylogenetic relationships tera) and other insects associated with Texas. Journal of Range Management 36, in the caesalpinioid legumes: a prelimi- Parkinsonia aculeata (Leguminosae) in 411-4. nary analysis based on morphological Argentina). Revista de Biologia Tropical Flanagan, G.J., van Rangelrooy, D.S. and and molecular data. In ‘Advances in 41, 463-70. Kerin, S. (1996). Integrated manage- legume systematics, Part 10’, eds B.B. Dimmit, M.A. (1987). The hybrid Palo ment of Parkinsonia aculeata on the Rop- Klitgaard and A. Bruneau, pp. 37-62. Verde ‘Desert Museum’: a new superior er River, Northern Territory, Australia. (Royal Botanic Gardens, Kew). tree for desert landscapes. Desert Plants Proceedings of the IX International Hughes, C.E. (1986). Progress in collection 8, 99-103. Symposium on Biological Control of and evaluation of NFT germplasm from 116 Plant Protection Quarterly Vol.24(3) 2009 Central America. Nitrogen Fixing Tree Le Houerou, H.N. (1984). Forage and fuel Miller, P. (1739). ‘The gardeners diction- Research Reports 4, 49-51. plants of N. Africa, near and middle ary’, 3rd edition. (London). Hughes, C.E. (1989). Intensive study of East. In ‘Plants for arid lands’, ed. G.E. Mohnot, K. and Chatterji, U.N. (1965). multipurpose tree genetic resources. Wickens, pp. 117-39. (George Allen and Chemico-physiological studies on the Unpublished report, Oxford Forestry Unwin, London). imbibition and germination of seeds Institute, UK. Lewis, G.P., Simpson, B.B. and Neff, J.L. of Parkinsonia aculeata Linn. Oester- Hughes, C.E. and Styles, B.T. (1984). Ex- (2000). Progress in understanding the reichische Botanische Zeitschrift 112, ploration and seed collection of mul- reproductive biology of the Caesalpin- 576-85. tiple-purpose dry zone trees in central ioideae (Leguminosae). In ‘Advances NWS (2000). National parkinsonia weed America. The International Tree Crops in legume systematics, part 9’, eds P.S. management strategy. (Queensland De- Journal 3, 1-31. Herendeen and A. Bruneau, pp. 65-78. partment of Natural Resources, Coor- Humphries, S.E., Groves, R.H. and Mitch- (Royal Botanic Gardens, Kew). paroo). ell, D.S. (1991). Plant invasions of Aus- Lockett, C., Gray, E. and Donnelly, G. Parsons, W.T. and Cuthbertson, E.G. tralian ecosystems: a status review and (1999). The prevalence of seed damage (2001). ‘Noxious weeds of Australia, management directions. In ‘Kowari 2’. by the introduced bruchid, Penthobru- second edition’. (CSIRO Publishing, (Australian National Parks and Wildlife chus germaini (Coleoptera: Bruchidae), Melbourne). Service, Canberra). a biological control agent for parkinso- PIER (Pacifi c Islands Ecosystems at Risk) Hunter, I.R. and Stewart, J.L. (1993). Fo- nia (Parkinsonia aculeata), in far north (2000). Invasive plant species: Parkin- liar nutrient and nutritive content of and central Queensland. Proceedings of sonia aculeata. www.hear.org/pier. Ac- Central American multipurpose tree the 12th Australian Weeds Conference, cessed March 2008. species growing at Comayagua, Hon- eds A.C. Bishop, M. Boersma and C.D. Plumier, L. (1703). ‘Nova plantarum Amer- duras. Commonwealth Forestry Review Barnes, pp. 431-5. (Tasmanian Weeds icanum genera’. (Regis and Regiae Sci- 72, 193-7. Society, Hobart). entiarum Academiae Typographum, Jones, C.E.Jr. and Buchmann, S.L. (1974). Lukitsch, B. and Wilson, A. (1999). Distri- Paris). Ultraviolet fl oral patterns as functional bution and impact of the mature seed Polhill, R.M. and Vidal, J.E. (1981). Caes- orientation cues in hymenopterous pol- feeding bruchid, Penthobruchus ger- alpineae. In ‘Advances in legume sys- lination systems. Animal Behaviour 22, maini on Parkinsonia aculeata in north- tematics’, Volume 1, ed. R.M. Polhill, 481-5. ern Australia. Proceedings of the 12th P.H. Raven, pp. 81-95. (Royal Botanic Julien, J.H. and Griffiths, M.W. (1998). Australian Weeds Conference, eds A.C. Gardens, Kew). ‘Biological control of weeds: a world Bishop, M. Boersma and C.D. Barnes, Rajaram, N. and Janardhanan, K. (1991). catalogue of agents and their target pp. 436-40. (Tasmanian Weeds Society, Studies on the underexploited tree weeds’, 4th edition. (CABI Publishing, Hobart). pulses, Acacia catechu Willd., Parkinso- Wallingford). MacBride, J.F. (1943). ‘Flora of Peru. Fiel- nia aculeata L. and Prosopis chilensis (Mo- Landsberg, R., Vitelli, M., Shepherd, B., diana 13 (3)’. (Field Museum of Natural lina) Stunz: chemical composition and Jayo, R. and Whalan, R. (2006). A catch- History, Chicago). antinutrutional factors. Food Chemistry ment approach to parkinsonia control – MacDicken, K.G. and Brewbaker, J.L. 43, 265-73. can it be done? Proceedings of the 15th (1984). Descriptive summaries of eco- Robertson, I.A.D. (1988). ‘A survey of the Australian Weeds Conference, eds C. nomically important nitrogen fixing natural enemies of Parkinsonia aculeata Preston, J.H. Watts and N.D. Crossman, trees. Nitrogen Fixing Tree Research Re- in Kenya’, 4 pp. Unpublished report of p. 481-3. (Weed Management Society of ports 2, 46-54. CABI Kenya Station.(CABI, Nairobi). South Australia, Adelaide). McKenzie, J.R., Pattison, M.J., Campbell, Robertson, K.R. and Lee, T. (1976). The Lawes, R.A. (2004). Temporal fl uctuations S.D. and Vitelli, J.S. (2006). The effect genera of Caesalpiniodeae in the south- of Parkinsonia aculeata L. seedling ger- camels have on ability of parkinsonia western United States. Journal of the Ar- mination and growth in three Austral- (Parkinsonia aculeata L.) to set seed. Pro- nold Arboretum 57, 1-54. ian rangelands habitats. Proceedings of ceedings of the 15th Australian Weeds Ross, R.H. (1998). Caesalpiniaceae: Parkin- the 14th Australian Weeds Conference, Conference, eds C. Preston, J.H. Watts sonia. Flora of Australia 12, 69-70. eds B.M. Sindel and S.B. Johnson, pp. and N.D. Crossman, pp. 886-9. (Weed Scott, K.A. (2006). Seed coat morphology 164-7. (Weeds Society of New South Management Society of South Austral- of Parkinsonia aculeata L. and pattern of Wales, Sydney). ia, Adelaide). heat-induced fractures. Proceedings of Lawes, R.A. and Grice, A.C. (2007). Con- McKenzie, J.R., Pattison, M.J., Steele, K.E., the 15th Australian Weeds Conference, trolling infestations of Parkinsonia ac- Campbell, S.D. and Vitelli, J.S. (2004). eds C. Preston, J.H. Watts and N.D. uleata in a riparian zone at the landscape Controlling dense infestations of par- Crossman, pp. 268-9. (Weed Manage- scale. Austral Ecology 32, 287-93. kinsonia (Parkinsonia aculeata L.). Pro- ment Society of South Australia, Ad- Lawes, R.A., Whiteman, L.V. and Grice, ceedings of the 14th Australian Weeds elaide). A.C. (2003). Mapping the distribution Conference, eds B.M. Sindel and S.B. Scrifres, C.J. (1980). ‘Brush management. of Parkinsonia aculeata in the Cape River Johnson, pp. 173-6. (Weed Society of principles and practices for Texas and catchment in northern Queensland, New South Wales, Sydney). the southwest’. (Texas A&M Press, Col- Australia. Plant Protection Quarterly 18, Madany, M.H. (1991). Living fences: lege Station). 152-6. Somali farmers adopt an agrofor- Sheikh, M.I., Shah, B.H. and Aleem, A. Lawson, B.E., van Klinken, R.D., Day, estry technology. Agroforestry Today 3, (1985). Effect of rainwater harvesting M.D., Heard, T.A. and Zalucki, M.P. 4-7. methods on the establishment of tree (2008). Improving the agent selection, Martin, T. and van Klinken, R.D. (2006). species. Forest Ecology and Management release and evaluation process: the role Value for money?: investment in weed 8, 257-64. of bioclimatic modelling. Proceedings management in Australian rangelands. Sprent, J.I. and Sutherland, J.M. (1990). Ni- of the 16th Australian Weeds Confer- Rangeland Journal 28, 63-75. trogen fi xing woody legumes. Nitrogen ence, eds R.D. van Klinken, V.A. Os- Miller, I.L. and Pickering, S.E. (1980). Fixing Tree Research Reports 8, 17-23. ten, F.D. Panetta and J.C. Scanlan, pp. Parkinsonia, a noxious weed. Agnote. Stewart, J.L., Dunsdon, A.J., Hellin, J.J. and 510-2. (Queensland Weeds Society, Bris- (Northern Territory Department of Pri- Hughes, C.E. (1992). Wood biomass es- bane). mary Production, Darwin). timation of Central American dry zone Plant Protection Quarterly Vol.24(3) 2009 117 trees. Oxford Forestry Institute Tropical van Klinken, R.D., Seabrook, L., Smith, C., Wyatt, R. (1981). Components of repro- Forestry Papers 26, Oxford. Ryan, J. and McAlpine, C. (2008). Us- ductive output in fi ve tropical legumes. Streets, R.J. (1962). ‘Exotic forest trees in ing spatial models and a Bayesian belief Bulletin of the Torrey Botanical Club 108, the British Commonwealth’. (Claren- network to predict risk of parkinsonia 67-75. don Press, Oxford). invasion at a landscape scale in north- Swarbrick, J.T. (1997). Weeds of the Pacifi c ern Australia. Proceedings of the 16th Islands. Technical paper No. 209. South Australian Weeds Conference, eds R.D. Pacific Commission, Noumea, New van Klinken, V.A. Osten, F.D. Panetta Caledonia. 124 pp. and J.C. Scanlan, p. 194. (Queensland Teketay, D. (1996). Germination ecology Weeds Society, Brisbane). of twelve indigenous and eight exotic Vitelli, J.S. (1995). Parkinsonia. In ‘Exotic multipurpose leguminous species from weeds and their control in north west Ethiopia. Forest Ecology and Management Queensland’, ed. N. March, p. 34-5 (De- 80, 209-23. partment of Lands, Brisbane). Thorpe, J.R. and Lynch, R. (2000). ‘The Vitelli, J.S. (2000). Options for effective determination of Weeds of National weed management. Tropical Grasslands Signifi cance’. (National Weeds Strategy 34, 280-94. Executive Committee, Launceston). Vitelli, J.S. and Madigan, B.A. (2004). Toh, R., Galea, V., Diplock, N. and van Evaluation of a hand-held burner for Klinken, R.D. (2008). Evaluation of the control of woody weeds by fl aming. fungal isolates for potential use as my- Australian Journal of Experimental Agri- coherbicides for seedbank reduction culture 44, 75-81. of Parkinsonia aculeata. Proceedings of Vitelli, J.S. and Pitt, J.L. (2006). Assessment the 16th Australian Weeds Conference, of current weed control methods rel- eds R.D. van Klinken, V.A. Osten, F.D. evant to the management of the biodi- Panetta and J.C. Scanlan, pp. 284-6. versity of Australian Rangelands. The (Queensland Weeds Society, Brisbane). Rangeland Journal 28, 37-46. Trama, F. (2004). Restauración de humedal von Mueller, F. (1876). ‘Select plants read- Palo Verde para aves. Revista ambientico, ily eligible for industrial culture or nat- no. 129. Online resource: www.ambi- uralisation in Victoria, with indications entico.una.ac.cr/129/index.htm. Ac- of their native countries and some of cessed March 2008. their uses’. (M’Carron, Bird and Co., Troup, R.S. and Joshi, H.B. (1983). The sil- Melbourne). viculture of Indian trees. Vol IC. Legu- Wagner, W.L., Herbst, D.R. and S.H. So- minoseae. (Controller of Publications, hmer (1999). ‘Manual of the fl owering Delhi). plants of Hawai’i’. 2 volumes. Bishop van Klinken, R.D. (2005). Total annual seed Museum Special Publication 83. (Uni- loss on a perennial legume through versity of Hawai’i and Bishop Museum predation by insects: the importance of Press, Honolulu) within-season seed and seed-feeder dy- Wiggins, I.L. and Porter, D.M. (1971). ‘Flo- namics. Austral Ecology 30, 414-25. ra of the Galapagos Islands’. (Stanford van Klinken, R.D. (2006). Parkinsonia bio- University Press, California). control: what are we trying to achieve? Wilson, C.G. and Miller, I.L. (1987). Par- Australian Journal of Entomology 45, 268- kinsonia in the Northern Territory. 71. Northern Territory Department of Pri- van Klinken, R.D. and Campbell, S. (2001). mary Industry and Fisheries Technical Biology of Australian weeds. 37. Pro- Bulletin. No. 106. sopis L. species. Plant Protection Quar- Woods, W. (1985). Bruchid seed beetles terly 16, 2-20. for control of Parkinsonia aculeata in van Klinken, R.D. and Flack. L. (2005). Australia. Proceedings of the VI In- The relationship between wet heat and ternational Symposium on Biological hard-seeded dormancy and germina- Control of Weeds, ed. E. Delfosse, pp. tion. Weed Science 53, 663-9. 855-62. (Agriculture Canada, Ottawa). van Klinken, R.D. and Flack, L. (2008). Woods, W. (1986). Biological control of What limits predation rates by a spe- parkinsonia. Western Australian Journal cialist seed feeder? Journal of Applied of Agriculture 27, 80-3. Ecology 45, 1600-11. Woods, W. (1988). The potential for bio- van Klinken, R.D., Flack, L. and Pettit, W. logical control of Parkinsonia aculeata L.: (2006). Wet-season dormancy release phytophagous insects collected from in seed banks of a tropical leguminous the USA, Mexico and Costa Rica and the shrub is determined by wet heat. Annals effect of insect damage on the growth of Botany 98, 875-83. and survival of the plant. M.Sc. thesis, van Klinken, R.D., Lukitsch, B. and Cook, University of Western Australia, Perth. C. (2008). Interaction between seed Woods, W. (1992). Phytophagous insects dormancy-release mechanism, envi- collected from Parkinsonia aculeata ronment and seed bank strategy for a (Leguminoseae: Caesalpiniaceae) in widely distributed perennial legume the Sonoran Desert Region of the south (Parkinsonia aculeata, Caesalpinaceae). western United States and Mexico. En- Annals of Botany. 102, 255-64. tomophaga 37, 465-74.