The current range of J. gossypiifolia includes Australia, Africa, Asia, North and South America. In Australia J. gossypiifolia is declared invasive in Queensland (2002)...
Bellyache bush invasion of a riparian site, North Queensland, Australia.
Copyright
Faiz Bebawi
Invasive habit
Bellyache bush invasion of a riparian site, North Queensland, Australia.
Faiz Bebawi
Title
Invasive habit
Caption
Dense infestation of an ephemeral river. Daly River, Northern Territory, Australia.
Copyright
Faiz Bebawi
Invasive habit
Dense infestation of an ephemeral river. Daly River, Northern Territory, Australia.
Faiz Bebawi
Title
Invasive habit
Caption
Grazed pasture invasion by bellyache bush. Australia.
Copyright
Faiz Bebawi
Invasive habit
Grazed pasture invasion by bellyache bush. Australia.
Faiz Bebawi
Title
Queensland bronze biotype
Caption
Queensland bronze bellyache bush biotype. Australia.
Copyright
Faiz Bebawi
Queensland bronze biotype
Queensland bronze bellyache bush biotype. Australia.
Faiz Bebawi
Title
Queensland green leaf biotype
Caption
Queensland green leaf bellyache bush biotype. Australia.
Copyright
Faiz Bebawi
Queensland green leaf biotype
Queensland green leaf bellyache bush biotype. Australia.
Faiz Bebawi
Title
Queensland purple leaf biotype
Caption
Queensland purple leaf bellyache bush biotype. Australia.
Copyright
Faiz Bebawi
Queensland purple leaf biotype
Queensland purple leaf bellyache bush biotype. Australia.
Faiz Bebawi
Title
Katherine green leaf biotype
Caption
Katherine green leaf bellyache bush biotype. Northern Territory, Australia.
Copyright
Faiz Bebawi
Katherine green leaf biotype
Katherine green leaf bellyache bush biotype. Northern Territory, Australia.
Faiz Bebawi
Title
Darwin purple leaf biotype
Caption
Darwin purple leaf bellyache bush biotype. Northern Territory, Australia.
Copyright
Faiz Bebawi
Darwin purple leaf biotype
Darwin purple leaf bellyache bush biotype. Northern Territory, Australia.
Faiz Bebawi
Title
Western Australia maroon leaf biotype
Caption
Western Australia maroon leaf bellyache bush biotype. Tipperary Station in the Northern Territory, Australia. Picture courtesy of Colin Wilson/Monitoring & Evaluation Project Officer, KI NRM, South Australia. http://www.kinrm.sa.gov.au/Home.aspx
Western Australia maroon leaf bellyache bush biotype. Tipperary Station in the Northern Territory, Australia. Picture courtesy of Colin Wilson/Monitoring & Evaluation Project Officer, KI NRM, South Australia. http://www.kinrm.sa.gov.au/Home.aspx
Western Australia green leaf bellyache bush biotype.
Copyright
Tracey Vinnicombe/Department of Agriculture and Food, Western
Western Australia green leaf biotype.
Western Australia green leaf bellyache bush biotype.
Tracey Vinnicombe/Department of Agriculture and Food, Western
Title
Scarlet leaf biotype
Caption
Scarlet leaf bellyache bush biotype. West Timor, Indonesia. Picture courtesy of Colin Wilson/Monitoring & Evaluation Project Officer, KI NRM, South Australia. http://www.kinrm.sa.gov.au/Home.aspx
Scarlet leaf bellyache bush biotype. West Timor, Indonesia. Picture courtesy of Colin Wilson/Monitoring & Evaluation Project Officer, KI NRM, South Australia. http://www.kinrm.sa.gov.au/Home.aspx
Female flower (a) and flower with male and female parts (b). Picture courtesy of Di Taylor/Alan Fletcher Research Station, Brisbane Australia. http://www.dpi.qld.gov.au/4790_11831.htm
Copyright
Di Taylor/Alan Fletcher Research Station, Brisbane, Australia
Flowers
Female flower (a) and flower with male and female parts (b). Picture courtesy of Di Taylor/Alan Fletcher Research Station, Brisbane Australia. http://www.dpi.qld.gov.au/4790_11831.htm
Di Taylor/Alan Fletcher Research Station, Brisbane, Australia
Title
Queensland bronze leaf and flower
Caption
Queensland bronze leaf biotype leaf and flower. Australia.
Copyright
Faiz Bebawi
Queensland bronze leaf and flower
Queensland bronze leaf biotype leaf and flower. Australia.
Faiz Bebawi
Title
Darwin purple fruits and flowers
Caption
Darwin purple leaf biotype fruits and flowers. Northern Territory, Australia.
Copyright
Faiz Bebawi
Darwin purple fruits and flowers
Darwin purple leaf biotype fruits and flowers. Northern Territory, Australia.
Faiz Bebawi
Title
Queensland green leaf biotype flowers and fruits
Caption
Queensland green leaf biotype showing flowers and fruits. Australia.
Copyright
Faiz Bebawi
Queensland green leaf biotype flowers and fruits
Queensland green leaf biotype showing flowers and fruits. Australia.
Faiz Bebawi
Title
Katherine green leaf biotype flowers
Caption
Katherine green leaf biotype flowers. Northern Territory, Australia.
Copyright
Faiz Bebawi
Katherine green leaf biotype flowers
Katherine green leaf biotype flowers. Northern Territory, Australia.
Faiz Bebawi
Title
Germinated seeds
Caption
Germinated seeds of bellyache bush. Note scale.
Copyright
Faiz Bebawi
Germinated seeds
Germinated seeds of bellyache bush. Note scale.
Faiz Bebawi
Title
Dissected seed
Caption
Dissected seed showing crustaceous testa and fleshy exotegmen.
Copyright
Faiz Bebawi
Dissected seed
Dissected seed showing crustaceous testa and fleshy exotegmen.
Faiz Bebawi
Title
Seed types
Caption
Bellyache bush seed types produced by Queensland bronze and Katherine green biotypes.
Copyright
Faiz Bebawi
Seed types
Bellyache bush seed types produced by Queensland bronze and Katherine green biotypes.
Faiz Bebawi
Title
Meat ants dispersing seeds
Caption
Meat ants dispersing bellyache bush seeds.
Copyright
Faiz Bebawi
Meat ants dispersing seeds
Meat ants dispersing bellyache bush seeds.
Faiz Bebawi
Title
Ant-chewed and intact seeds
Caption
Ant-chewed and intact bellyache bush seeds.
Copyright
Faiz Bebawi
Ant-chewed and intact seeds
Ant-chewed and intact bellyache bush seeds.
Faiz Bebawi
Title
Bellyache bush seedlings
Caption
Bellyache bush seedlings emerging from seeds deposited in ant middens.
Copyright
Faiz Bebawi
Bellyache bush seedlings
Bellyache bush seedlings emerging from seeds deposited in ant middens.
Faiz Bebawi
Title
Shallow root system
Caption
Bellyache bush showing shallow root system.
Copyright
Faiz Bebawi
Shallow root system
Bellyache bush showing shallow root system.
Faiz Bebawi
Title
Defoliated bellyache bush plants
Caption
In Australia, the bellyache bush loses most of its leaves during the dry season.
Copyright
Faiz Bebawi
Defoliated bellyache bush plants
In Australia, the bellyache bush loses most of its leaves during the dry season.
Faiz Bebawi
Title
Natural enemy
Caption
Jewel bug (Agonosoma trilineatum: Hemiptera) - female left, male right - feeding on a bellyache bush pod.
Copyright
Faiz Bebawi
Natural enemy
Jewel bug (Agonosoma trilineatum: Hemiptera) - female left, male right - feeding on a bellyache bush pod.
Faiz Bebawi
Title
Australian distribution map
Caption
Bellyache bush Australian distribution map. (Map courtesy of Marc Bryant/Project Manager BioSIRT-Biosecurity Queensland)
Copyright
Marc Bryant/Project Manager BioSIRT-Biosecurity Queensland
Australian distribution map
Bellyache bush Australian distribution map. (Map courtesy of Marc Bryant/Project Manager BioSIRT-Biosecurity Queensland)
Marc Bryant/Project Manager BioSIRT-Biosecurity Queensland
The current range of J. gossypiifolia includes Australia, Africa, Asia, North and South America. In Australia J. gossypiifolia is declared invasive in Queensland (2002), the Northern Territory (2001) and Western Australia (1976) (Queensland Government, 2003; Randall et al., 2009). The presence of several morphological, physiological and reproductive attributes including sympodial growth habit, a shallow root system, an abundance of extra-floral nectaries (provides more incentive for ant visitation), extreme adaptation to xeric (short stems) and non-xeric (temporary flooding) conditions coupled with high stem sugar reserves and sap content (physiological), prolific production of long-lived myrmecochorous seeds (ant-attractant substances) that have the ability to float in water, coupled with the ability to reproduce vegetatively and to prolifically flower and set seed all year, and absence of natural enemies, all combined together, contribute to its successful invasive strategy (Bebawi and Campbell, 2002a, b, c, d). J. gossypiifolia also appears to exhibit allelopathy, which in addition to its choking growth and ability to form dense monocultures imposes fundamental changes to fire regimes, community composition, structure, and function.
The genus name Jatropha combines the Greek iatros, meaning physician, with tropheia, meaning mother’s milk, hinting at the medicinal properties of the plant (Parsons and Cuthbertson, 2001). The species name ‘gossypiifolia’ is a combination of the Latin gossypium, meaning cotton, and folium, suggesting that the leaves appear similar to those of the cotton plant (Parsons and Cuthbertson, 2001).
The genus Jatropha belongs to the tribe Jatrophieae of Crotonoideae in the family Euphorbiaceae and the genus contains approximately 186 species (Govaerts et al., 2000). Dehgan and Webster (1979) divided the genus into two subgenera (Curcas and Jatropha) with 10 sections and 10 subsections. They postulated that physic nut (Jatropha curcas L.) is the most primitive form of the genus and that J. gossypiifolia evolved from the physic nut. The facultative annual growth habit (apical dominance absent) of bellyache bush is considered a more phylogenetically advanced growth habit than the arborescent growth habit (apical dominance present) of physic nut (Dehgan and Webster, 1979). A taxonomic characteristic of the genus Jatropha is the occurrence of either latex-cells or latex vessels (Rao and Malaviya, 1964).
Although most Jatropha species are native to the New World (the Americas), no complete revision of the Old World (Europe, Asia, and Africa) Jatropha exists (Heller, 1996). Hemming and Radcliffe-Smith (1987) revised 25 Somalian species, all of the subgenus Jatropha, and placed them in six sections and five subsections.
Different varieties of bellyache bush exist in several countries (Backer and van der Brink, 1963; Dehgan, 1982; Sreenivasa Rao and Raju, 1994). For example, J. gossypiifolia L. var. elegans was listed in the flora of Java (Backer and van der Brink, 1963) and three varieties, J. gossypiifolia var. elegans, J. gossypiifolia var. gossypiifolia and J. gossypiifolia var. staphysagrifolia (Mill.) Müll. were identified in the United States (Dehgan, 1982). Sreenivasa Rao and Raju (1994) reported J. gossypiifolia L. (var. gossypiifolia and var. (Pohl) Müll.Arg.) from India.
In Australia, several biotypes have been noted based on morphological, phenological, and physiological differences (Pitt and Miller, 1991; Bebawi and Campbell, 2004; Bebawi et al., 2007e, 2009). Detailed taxonomic, genetic and ecological studies are now required to verify the differences which could have implications for management, particularly selection of biological control agents. Such studies would also help determine whether there is only one variety present or if some of the noted biotypes could in fact be different varieties. It has been suggested that two of the biotypes in Australia could potentially be J. gossypiifolia var. elegans and J. gossypiifolia var. staphysagrifolia (B Dehgan, University of Florida, USA, personal communication, 2008). A recent molecular genetic study identified that multiple introductions of diverse haplotypes from throughout the native range has occurred in Australia and may explain some of the considerable variation that is found between infestations (Prentis et al., 2008).
J. gossypiifolia is an erect, woody, deciduous, tropical or sub-tropical perennial. Plants commonly grow between 2 to 3 m high (Parsons and Cuthbertson, 2001). In Queensland (Australia), plants are capable of growing to 4 m with a canopy diameter of 2 m and a basal stem diameter up to 15 cm (Bebawi and Campbell, 2002b; Vitelli and Madigan, 2004). Heights over 3 m have also been reported in shaded areas in the Northern Territory (Pitt and Miller, 1991). In its native habitat, for example, in Puerto Rico J. gossypiifolia plant height may range between 0.5–2.0 m high and 1–3 cm in basal stem diameter (Francis, 2004). In Australia, J. gossypiifolia is capable of attaining a height of up to 2 m in a single growing season (Dehgan and Webster, 1979), particularly in areas free from competition, such as riverbeds. In a competition trial in the dry tropics of north Queensland, J. gossypiifolia grown on pasture-cleared plots reached average heights of 64 cm and 113 cm after the first and second year of establishment respectively (FF Bebawi, Biosecurity Queensland, Australia, personal communication, 2009). Similarly, in Puerto Rico J. gossypiifolia was found to grow up to 0.5 m per year (Liogier, 1990).
The stem habit is sympodial and apical dominance is absent because branching occurs at reproductive maturity and thereafter at subsequent flowering episodes. Following germination, the primary stem (main axis) grows until flowering is initiated (FF Bebawi, Biosecurity Queensland, Australia, personal communication, 2009). The stem then starts forking out producing axillary shoots (stems) and more flowers. Stems are green when young, but invariably turn ‘grey’ with age and extremely sticky due to the presence of extra-floral nectaries.
Leaves are arranged alternately along the stem and may be bright green, dark green, bright or dark bronze, bright red or bright purplish-red depending on biotype, variety or leaf maturity (Pitt and Miller, 1991; Burkill, 1994; Parsons and Cuthbertson, 2001; Bebawi et al., 2005b). Leaf petioles are 2–7 cm long and the leaf blades are palmately 3–5 lobed, generally 45–90 × 50–130 mm in size, and more or less elliptic (Wheeler et al., 1992). The lamina is glabrous and the leaf margin is denticulate with venations ending in stipitate glandular hairs.
The root system is relatively small compared with the shoot system, with stem to root ratios increasing with age. The stem to root dry weight biomass ratio of juvenile, mature, and adult J. gossypiifolia plants in north Queensland averaged 5.6, 6.1 and 7.1 to 1, respectively (FF Bebawi, Biosecurity Queensland, Australia, personal communication, 2009). J. gossypiifolia has a fleshy shallow root system with four short robust lateral roots and many fine tertiary roots (Singh, 1970; Howard, 1989; Liogier, 1990; Burkill, 1994; Csurhes, 1999; Parrotta, 2001). Fresh root weight of a dense infestation (20 000 plants ha-1) of juvenile (up to 20 cm height), mature (20–100 cm height) and adult (>100 cm height) J. gossypiifolia plants in north Queensland averaged 0.5, 2.8 and 10.5 t ha-1 (25, 137, and 523 g plant-1), respectively. Root moisture content of juvenile, mature, and adult J. gossypiifolia plants is relatively similar, averaging 81%, 73%, and 71%, respectively (FF Bebawi, Biosecurity Queensland, Australia, personal communication, 2009).
Large variation in flower colour can occur. For example, in Australia the small flowers range from light red to dark purple on the outer sections of petals, with dull to bright yellow centres (FF Bebawi, Biosecurity Queensland, Australia, personal communication, 2009). Male flowers are cup-shaped with a diameter of 6–9 mm. There are eight stamens all fused into a tube like structure, connate at the base, the upper portion is free with two tiers of anthers of unequal length (Reddi and Reddi, 1983). There are three relatively large anthers in the upper tier and five relatively small in the lower. The anthers are dorsifixed. Pollen grains are spherical, bright yellow with a sticky, oily coating (Reddi and Reddi, 1983). Female flowers are quite similar in shape to male flowers but they are larger, with a diameter of up to 9 mm. Sepals as well as petals are larger than those on the male. Styles are 3–4, slender dilated into a capitate, bifid stigma. The stigmas are pale green and very sticky (Reddi and Reddi, 1983). Inflorescences are glandular and hairy (de Padua et al., 1999). Flowers are pedicellate, terminal and occur in corymbose cymes. Flower bracts are linear-lanceolate with glandular margins (Dehgan and Webster, 1979). Flowers are radially symmetrical and loaded with nectaries.
The calyx consists of five united sepals, which are ovate to lanceolate, acuminate with stalked capitate glands on the margins, distinctly imbricate and 4 mm long. The corolla consists of five petals which are obovate and 3–5 mm long. Sepals touch at the base to form a short green tube and have stipitate marginal glands (FF Bebawi, Biosecurity Queensland, Australia, personal communication, 2009).
The fruit has three lobes (trilocular). It is an explosively dehiscent capsule with a single seed per locule and axile placentation. It is globose, pedicellate, generally bright green and woody at maturity, turning pale green or tan when ripe (Berg, 1975; Dehgan and Webster, 1979; Burkill, 1994; Parrotta, 2001). Most fruits bear three seeds.
When dissected, seeds have a thin, outer whitish-fleshy envelope (exotegmen) that surrounds a thin and shell-like testa. A longitudinal section of mature seeds show an exotegmen connected to the caruncle at the micropylar region (proximal end) of the seed by dense parenchymatous tissue. When the seed ripens the surrounding exotegmen fuses to the seed tegmen. The ‘exotegmen’ is fully transparent, shiny, and slippery when mature, and securely fused to the testa. The testa tegmen is below the ‘exotegmen’. When the testa is removed, a white mass called the kernel is left. The kernel is enclosed in a very thin membrane (endotegmen). The kernel consists predominantly of a thick, soft, and oily endosperm (Singh, 1970; Bebawi and Campbell, 2004) that surrounds an embryo with two papery cotyledons in the centre. Seed germination is epigeal, i.e. the cotyledons appear above the ground surface.
Latex cells occur in the stem, leaf, petiole, flower parts, fruit, and in the seed coat (Rao and Malaviya, 1964). Latex from the shoot apex is transparent to nearly white changing to reddish brown on exposure to air (FF Bebawi, Biosecurity Queensland, Australia, personal communication, 2009). Dehgan and Webster (1979) also recognized four extra floral gland systems in the lamina, petiole, stipules and bracts in the genus Jatropha. The bark is green, succulent, and smooth.
Plant Type
Top of page
Broadleaved Perennial Seed propagated Shrub Vegetatively propagated Woody
J. gossypiifolia is native to the New World but has been cultivated widely in tropical countries (Burkill, 1994; de Padua et al., 1999). In North America it is found mainly in Florida and in Mexico (Francis, 2005). In South America it occurs predominantly in Bolivia, Venezuela, Ecuador, Peru, and Brazil (de Padua et al., 1999). It is also a common plant in the Caribbean, mainly in the Dominican Republic, Puerto Rico and Leeward Islands, as well as in the Hawaiian Islands (Csurhes, 1999; de Padua et al., 1999).
Introduced to southern Africa, the plant has spread from Mozambique through Zambia to the Transvaal and Natal. In West Africa, it is listed in the exotic flora of Chad (Brundu and Camarda, 2004), Cameroon and Ghana (Csurhes, 1999).
In Australia, J. gossypiifolia is found in the northern half of the continent (Ashley, 1995). In northern Queensland, it is present in riparian and sub-riparian habitats of the Burdekin, Walsh, Palmer, Flinders, and Gregory Rivers and the headwaters of Lake Eyre Basin. Small scattered infestations occur in central Queensland, particularly in the Fitzroy catchment (Csurhes, 1999; Barron, 2004). In the Northern Territory, J. gossypiifolia persists in several coastal areas, but is most prevalent further inland (Csurhes, 1999). Among the worst infestations in the Northern Territory are those in the Darwin area (Channel Island), the Daly River catchment, the Gulf of Carpentaria region and the Barkly Tablelands. Other infestations occur in the McArthur River, Roper River, and Victoria River catchments (Miller and Pitt, 1992). In Western Australia, widespread populations occur in the east Kimberley area, with larger infestations in the Lake Argyle catchment and near Halls Creek (Parsons and Cuthbertson, 2001; Anon., 2005). Small controlled infestations occur in the west Kimberly region at the De Grey River, Broome, and Fitzroy River and on the Drysdale River and Koolan Island in the north Kimberly (King and Wirf, 2005; Turpin et al., 2006). The potential range of J. gossypiifolia in Australia includes the entire tropical savannas (Thorp and Lynch, 2000).
The distribution in this summary table is based on all the information available. When several references are cited, they may give conflicting information on the status. Further details may be available for individual references in the Distribution Table Details section which can be selected by going to Generate Report.
Alice Springs, Daly River Catchment, Gulf of Carpentaria, Barkly Tablelands, McArthur, Roper, Victoria River Catchments, Kakadu and Elsey National Park, Flora River Conservation Reserve, Leper Island, Channel Island, Tipperary Station, Katherine, Pine Creek, Timber Creek, Borroloola, Nhulunbuy, Elliot, Mataranka, Middle Point, Batchelor, Aroona Creek System
East Kimberley, Lake Argyle Catchment, Halls Creek, West Kimberley region, De Grey, Broome, Fitzroy, Drysdale Rivers, Turkey Creek, Bow River, Kununurra, Wyndham, Mabels Down Station, Derby, Koolan Island
J. gossypiifolia is a plant widely exploited traditionally in Latin America and the Caribbean, with references to its use for medicinal purposes dating back to the 1800s (Morton, 1981). It is also widely distributed throughout the tropics as an ornamental, and was probably distributed by Portuguese seafarers via Cape Verde and Guinea Bissau to other countries in Africa and Asia.
J. gossypiifolia was introduced to Australia in the latter part of the 1800s (Everist, 1974; Pitt and Miller, 1991; Parsons and Cuthbertson, 2001). It was introduced to the Northern Territory in 1888 along with other Jatropha species, probably for medicinal and ornamental purposes (Anon., 1888; Pitt and Miller, 1991; Pitt, 1997) and was listed as a naturalized component of Queensland’s flora in 1912 (Bailey, 1912).
The potential for further spread of J. gossypiifolia in Australia and other countries in the Asia and Pacific region is very high particularly as global warming gathers momentum and more land is deforested for agricultural purposes coupled with the need for medicinal and bio-fuel plants. Global warming is expected to shift current favourable ecological latitudinal zones further south in Australia, Asia and South America. For example, higher temperatures and wetter conditions across central Australia will be the driving ecological factor that could promote successful establishment of J. gossypiifolia not only from seed spread but also from flood dislodged plants. Increases in land clearance for agricultural purposes could also facilitate quick establishment of J. gossypiifolia.
Many Jatropha species in their native Americas occur in seasonally dry areas such as grassland-savanna (cerrado), thorn forest scrub and caatingas, and are completely lacking from the moist Amazon humid forest region (Dehgan and Schutzman, 1994). However, J. gossypiifolia is an opportunistic colonizer of disturbed habitats and is frequently found in areas where the natural vegetation has been over-grazed or removed by human activity or floodwaters (Csurhes, 1999). In Java (Backer and van den Brink, 1963) and Andhra Pradesh in India (Rao and Raju, 1994), it has become very common along roadsides, railway tracks and eroded places. Similarly, Csurhes (1999) remarked that J. gossypiifolia has become abundant in northern Australia along roadsides, around abandoned homesteads and near old mine sites, suggesting that colonization in such areas is probably an indicator of disturbance. In Timor it grows from upland country right down to the bed of rivers (Miller and Pitt, 1992).
J. gossypiifolia is particularly well adapted to a variety of habitats particularly the seasonally wet/dry climate of northern Australia (Csurhes, 1999), where it invades lowland habitats such as riparian zones, ephemeral watercourses, pastures and other disturbed areas. Riparian zones are most vulnerable to J. gossypiifolia invasion, possibly because fire often fails to penetrate riparian vegetation (Csurhes, 1999).
Most naturalized populations of J. gossypiifolia in Australia grow in areas receiving 400–1200 mm average annual rainfall, with some of the heaviest infestations found where average annual rainfall is 600–1000 mm (Csurhes, 1999). In Puerto Rico, it may be found in areas receiving from 750 to about 2000 mm of annual precipitation (Liogier, 1990). In wetter areas such as parts of the Northern Territory in Australia, J. gossypiifolia has shown a 146% expansion of its area per year (Pitt and Miller, 1991). However, even under dry conditions, J. gossypiifolia is capable of spreading. For example, on a property located in the Ravenswood area of northern Queensland, a J. gossypiifolia infestation increased in size by 76% over three years, despite the area receiving only half its average annual rainfall (Vogler and Keir, 2005).
J. gossypiifolia grows on nearly all types of soils within its range. In Australia, it grows equally well in saline and non-saline soils but seems to prefer sandy loams (Csurhes, 1999). It has been found growing down to the high tide mark near Darwin in the Northern Territory. In Puerto Rico, it is more common in soils with high base saturation, such as dry areas, sites near the ocean, and soils derived from limestone (Liogier, 1990).
J. gossypiifolia is intolerant of shade (Burkill, 1994). Although plants may survive for several seasons in moderate shade, they need full or nearly full sunlight for longer-term survival and fruiting (Burkill, 1994). J. gossypiifolia is very sensitive to frost, which damages the apices of stems (FF Bebawi, Biosecurity Queensland, Australia, personal communication, 2009). It may also suffer from water-logging in poorly drained soils (Dehgan, 1982) or as a result of flooding. In north Queensland, flooding of the Palmer River in March and April 2006 killed all plants that were growing in the riverbed (FF Bebawi, Biosecurity Queensland, Australia, personal communication, 2009).
J. gossypiifolia is diploid and the basic chromosome number is 11 (2n = 22) (Nanda, 1962; Datta, 1967; de Padua et al., 1999). A recent molecular genetic study identified that multiple introductions of diverse haplotypes from throughout the native range has occurred in Australia and may explain some of the considerable variation that is found between infestations (Prentis et al., 2008).
Reproductive Biology
J. gossypiifolia is capable of reaching reproductive maturity very quickly under favourable moisture conditions. In pot trials, plants flowered 55 days after germination (Bebawi et al., 2005c). In the field in north Queensland, the time to first flowering averaged 74 days in cleared areas, 294 days in rocky sites, and 454 days in grazed pastures (Bebawi et al., 2005c). It was suggested that more plants flowered in cleared areas and reached reproductive maturity earlier because of the absence of inter-specific competition (Bebawi et al., 2005c). Increasing J. gossypiifolia density also caused exponential increases in time to first flowering (Bebawi et al., 2005c). At the lowest density (20 plants m-2) plants flowered 55 days after germination. In contrast, only 6.3% of plants within the highest density (320 plants m-2) had flowered 42 months after germination (Bebawi et al., 2005c). In a pot trial, the percentage of J. gossypiifolia plants that flowered decreased exponentially with increased density (Bebawi et al., 2005c).
Once J. gossypiifolia plants reach reproductive maturity, they have a definite pattern of flowering within inflorescences. More female than male flowers mature on the first day, with female flowering showing a declining trend thereafter over a five-day period (Reddi and Reddi, 1983). In contrast, the number of male flowers gradually increases so that they dominate by the twelfth day, after which their frequency gradually declines. In effect, the inflorescence is protogynous (Reddi and Reddi, 1983). Dehgan and Webster (1979) showed that there was a lag of 24–48 h in anthesis of male flowers after opening of female flowers.
The ratio of male to female flowers is on average 11:1 (Reddi and Reddi, 1983; Wild, 2003). Dehgan (1983) found that short days result in production of more male flowers, while long days caused either a drastic increase in or total change to female flowers. Consequently, the ratio of male to female flowers and therefore seed production may be influenced by latitude and season.
Female flowers produce nearly 1.6 times more nectar than male flowers (Reddi and Reddi, 1983). A positive association has been detected between temperature and nectar concentration and an inverse one between relative humidity and nectar concentration (Reddi and Reddi, 1983).
J. gossypiifolia nectar consists of glucose, sucrose and fructose, amino acids and proteins (Reddi and Reddi, 1983). The nectar is attractive to insects, which are essential for normal seed set (Reddi and Reddi, 1983; Wild, 2003). Pollination may occur through selfing, because the flowers are self-compatible, or through outcrossing (Dehgan and Webster, 1979). The former mechanism ensures perpetuation of the species and the latter maintains species heterozygosity (Dehgan and Webster, 1979).
Many insect species have been observed foraging on nectar of J. gossypiifolia in Queensland (Australia) (Bebawi et al., 2007d). Most of the insects are beneficial to J. gossypiifolia because they perform vital roles such as pollination, defence and dispersal. For example, meat ants (Iridomyrmex spadius), honey bees (Apis mellifera) and oleander butterflies (Euploa core-corinna) visit flowers and perhaps pollinate J. gossypiifolia. Hives of the stingless bee Trigona carbonaria were successfully used to pollinate J. gossypiifolia grown in glasshouses.
Seed production
Seed production of J. gossypiifolia is generally prolific, but many factors such as environmental conditions, plant biotype/variety, plant density and location may influence the quantity of seeds produced. Adult plants growing in Queensland (Australia) have been found to produce between 2000 and 12 000 seeds plant-1y-1 (Bebawi and Campbell, 2002a), which corresponds to around 170 kg ha-1. In Western Australia J. gossypiifolia is reported to produce between 1800 and 2400 seeds plant-1y-1 (APB Infonote, 1994). Annual seed production of J. gossypiifolia in the Indian sub-continent was estimated at 500 kg ha-1 (Raina and Gaikwad, 1987).
Dense infestations of J. gossypiifolia growing in a relatively dry location in north Queensland produced four seeds m-2 during the 2003–2004 wet season (288 mm rainfall), in comparison to 343 seeds m-2 for plants growing in a wetter location (506 mm rainfall) (Vogler and Keir, 2005). Plants produce fewer seeds at high density. Above 40 plants m-2, seed production per plant began to decline. At very high densities (in excess of 300 plants m-2), there was very little seed production (Bebawi et al., 2005c).
Germination
Fresh J. gossypiifolia seeds exhibit high viability, but low germinability (Bebawi and Campbell, 2004). For example, fresh intact seed collected in north Queensland was 88% viable but only 10% of these were readily germinable (Bebawi and Campbell, 2004). Similarly, germination of seeds from Puerto Rico averaged just 4% (Liogier, 1990). Innate (primary) dormancy has been reported for other Euphorbiaceae (Ellis et al., 1985).
Seed type, seed weight, geographical location, temperature, control technique, and ants affect germination of J. gossypiifolia seed (Liogier, 1990; Bebawi and Campbell, 2002a, d, 2004). Seeds commence germination with the start of the wet season (Ashley, 1995) but will germinate at other times of the year if environmental conditions are favourable. For example, in north Queensland germination will often occur throughout the year on rainfall events that exceed 25 mm (FF Bebawi, Biosecurity Queensland, Australia, personal communication, 2009).
Some ants, (such as meat ants (Iridomyrmex spadius)) promote the germination of J. gossypiifolia seeds (Bebawi and Campbell, 2004). In a laboratory study 98% of ant-discarded seeds were viable and readily germinable (100%). Viability of intact seeds that had the caruncle manually removed was on average 10% lower than that of ant-discarded seeds, and only 8% of fresh viable intact seeds were germinable (Bebawi and Campbell, 2004). Removal of caruncles imitates the effect of meat ants on germination but not to the same extent as in ant-discarded seeds (Bebawi and Campbell, 2004). Additional scarification by ants on external seed structures may also promote germination.
Optimal germination temperatures for intact J. gossypiifolia seeds occur between 24 and 31°C. Germination generally commences five days after the imposition of favourable environmental conditions and reaches a maximum between days 11 and 12 (FF Bebawi, Biosecurity Queensland, Australia, personal communication, 2009).
Seed longevity
In a seed burial trial comparing germination and viability of intact and ant dispersed seeds exposed to either nil (rainfall excluded) or natural rainfall, no intact seeds exhumed after four years remained viable under natural rainfall conditions, whereas some ant-discarded seeds were still viable (3%). However, both intact and ant-discarded seeds exhumed after four years were 20% viable when rain was excluded (Bebawi et al., 2007b). Nil viability was first recorded at 12 months for intact seeds that were buried at 5-10 cm and exposed to natural rainfall conditions (Bebawi et al., 2007b). Seeds located on the surface generally persisted for longer than if buried at depth (Bebawi et al., 2007b). However, recent results from this trial showed that under rainfall excluded conditions, all intact seeds expired 84 months after burial whereas ant-discarded seeds showed some signs of viability (average 2%) 96 months after burial (DEEDI, 2009).
In another seed bank depletion trial, seedling emergence was still occurring after 51 months at a rocky site away from a river, whereas emergence had finished earlier (37 months) at a heavy clay soil site within the same period (FF Bebawi, Biosecurity Queensland, Australia, personal communication, 2009). Differences in seed bank depletion were attributed to differences in soil moisture conditions between the two sites.
Seedling establishment
In north Queensland, seedling densities can be very high under favourable environmental conditions. Averages of 247, 126, and 90 seedlings m-2 were measured within rocky, sub-riparian and riparian infestations of J. gossypiifolia, respectively (FF Bebawi, Biosecurity Queensland, Australia, personal communication, 2009). Even higher seedling densities (300–400 seedlings m-2) were recently recorded under dense canopies of J. gossypiifolia (Vogler and Keir, 2005).
Treatment of J. gossypiifolia infestations can result in massive recruitment of seedlings (Bebawi and Campbell, 2002c; Vitelli and Madigan, 2002; Bebawi et al., 2004). For example, J. gossypiifolia increased in density from five plants m-2 prior to aerial application of foliar herbicides to 400 plants m-2 post treatment (Vitelli and Madigan, 2002). Similarly, for every plant killed by foliar spraying, slashing, stick-raking and burning as part of an integrated research trial, 20, 97, 74 and 69 seedlings emerged, respectively (Bebawi et al., 2004). Treatments that caused the greatest soil disturbance appeared most conducive for seedling recruitment. Guterres et al. (2008) also found that burning promotes increased establishment of seedlings with subsequent rain.
High seedling mortality generally occurs, particularly if rainfall is limited and/or there is strong intra- or interspecific competition (Bebawi and Campbell, 2002c; Vitelli and Madigan, 2002). For example, in a fire trial seedling density in burnt plots reached a peak of 390 seedlings m-2 following favourable rainfall but declined to 30 seedlings m-2 one year later. Similarly in unburnt plots, seedlings peaked at 200 m-2 before crashing to 5 m-2 during the same period (Bebawi and Campbell, 2002c). A field trial that compared the impact of a range of four J. gossypiifolia seedling densities on seedling survival showed that 12 months after seedling emergence, seedling density was reduced by 75, 89, 98 and 99% at 25, 125, 250 and 500 seedlings m-2, respectively (FF Bebawi, Biosecurity Queensland, Australia, personal communication, 2009).
Even if high mortality of seedlings occurs under relatively dry conditions, sufficient recruitment for re-infestation of treated sites and expansion of infestations may occur in the absence of follow-up control activities. Observations by the authors suggest that once seedlings reach about 20 cm in size they are very hardy and will generally tolerate extreme environmental conditions.
Vegetative reproduction
J. gossypiifolia can readily regenerate from stem cuttings (e.g. dumped garden plant material) (Pitt and Miller, 1991; Parsons and Cuthbertson, 2001) and whole plants that may be removed during control activities or flood events. The shallow root system of J. gossypiifolia allows seedlings and mature plants to be easily dislodged (Pitt and Miller, 1991).
In Australia, several landholders have found J. gossypiifolia plants reshooting several months after they were pulled and left lying on the ground. Similarly, during a simulated slashing trial most off-cuts of J. gossypiifolia (particularly those cut during the dry season) flowered and produced capsules with viable seed up to 12 months after being cut (Bebawi and Campbell, 2002b), although a few of these had grown some new roots which were connecting them to the ground. Recently, Guterres et al. (2008) also found that over 50% of cut stems re-sprouted leaves during the dry season.
Physiology and Phenology
J. gossypiifolia utilizes the C3 photosynthetic pathway (Tezara et al., 1998; Fernandez et al., 1999; Bebawi and Campbell, 2002b). C3 plants are more efficient than C4 and CAM (Crassulacean Acid Metabolism) plants under cool and moist conditions and under normal light because they require fewer enzymes and no specialized anatomical adaptations (Rengifo et al., 2002).
Elevated CO2 had no significant effect on morning xylem water potential, leaf osmotic potential, or pressure potential of J. gossypiifolia grown under simulated seasonal drought conditions (Rengifo et al., 2002). Apparently, J. gossypiifolia has the capacity to respond to elevated CO2 and simulated drought by significantly reducing the proportional thickness of its leaf photosynthetic tissue systems (Rengifo et al. 2002) as well as by closing stomata (Tezara et al., 1998). This enhances water use efficiency and increases the photosynthetic rate. J. gossypiifolia may therefore exhibit increased biomass and growth rates, leading to further expansion of its current distribution, if atmospheric CO2 increases further.
In Australia, J. gossypiifolia loses most of its leaves during the dry season. Those that remain are generally quite small and concentrated at the apices of stems (Bebawi et al., 2005b). The transition between the two stages is fairly rapid—plants are either in full leaf or almost leafless (Bebawi et al., 2005b).
In north Queensland, an average of 17 and 20 leaves per stem have been recorded during the wet season (November–April) on plants growing within sub-riparian and riparian habitats, respectively (Bebawi et al., 2005b). On average there is one leaf per stem in the dry season (June–August). However, along watercourses and dam walls J. gossypiifolia plants may retain most of their leaves all year (Howard, 1989; Liogier, 1990; Burkill, 1994; Parrotta, 2001).
Flowering generally occurs from June to April in riparian zones and from September to April in sub-riparian zones of north Queensland (Bebawi et al., 2005b). While moisture availability appears to be a key driver of flower production, temperature may influence timing and duration (Bebawi et al., 2005b). In the Kimberley region of Western Australia, J. gossypiifolia generally flowers and fruits from February to May (Wheeler et al., 1992). In north Queensland, fruit production commences around September/October within riparian and sub-riparian zones and may last up to 10 months (Bebawi et al., 2005b). Flowering in India occurs from February through July (Parrotta, 2001).
Perennation
In Puerto Rico J. gossypiifolia is reported as an annual shrub (Ocampo and Balick, 2009). In other South American countries it is reported as a short-lived perennial shrub (Burkill, 1994). In north Queensland field trials indicated that plants can live for longer than 10 years (FF Bebawi, Biosecurity Queensland, Australia, personal communication, 2009), with anecdotal evidence suggesting greater than 20 years. However, there is a paucity of information on its longevity.
Seed dispersal occurs initially via dehiscent capsules that are capable of catapulting seeds as far as 13 m (Bebawi and Campbell, 2002a); although in dense infestations most seeds fall close to the parent plant. Once on the ground, some ants (particularly native meat ants (Iridomyrmex spadius)) disperse J. gossypiifolia seeds (Bebawi and Campbell, 2002a). In one study, an average of 12 330 seeds were retrieved from the middens (refuse piles) of individual meat ant nests over 12 months, with highest numbers recorded between February and June (Bebawi and Campbell, 2004). Meat ants appear to feed on the caruncle and exotegmen of J. gossypiifolia seeds and when finished discard the seed in their middens. The middens provide an improved environment (e.g. high nutrient status and absence of fire) for germination and survival of seedlings.
Water, humans and other animals, particularly the great bowerbird (Chlamydera nuchalis) are identified as potential long-distance dispersers of J. gossypiifolia seeds (APB Infonote, 1994; Ashley, 1995; Smith, 1995; FF Bebawi, Biosecurity Queensland, Australia, personal communication, 2009). Of these, floodwater is considered most important within catchments and humans are recognized as the main dispersers of J. gossypiifolia at a national and global scale (Csurhes, 1999).
Cattle grazing is one of the main economic land uses of tropical savannas in northern Australia. J. gossypiifolia is a vigorous invader of grazing lands where it can incur additional expenditure and threaten the viability of grazing enterprises. The economic impact of J. gossypiifolia on the livestock industry has been reported in Australia (Tothill et al., 1982), particularly in north Queensland, where direct losses to the pastoral industry occurred during drought due to poisoning of cattle, horses and goats (Csurhes, 1999). All cases of poisoning occurred in the dry season when pasture quality and quantity was at its lowest. Two landholders alone have spent approximately $56 240 on control of J. gossypiifolia in the Burdekin Catchment, north Queensland (Csurhes, 1999). A property owner within the Burdekin Catchment spent more than $50 000 over 10 years in an attempt to prevent J. gossypiifolia spreading along 17 km of river frontage (Bowen Independent, 1996). In the long term, J. gossypiifolia restricts the growth of native and introduced grasses, ultimately displacing pasture species, making any land colonised by the weed entirely unproductive (King and Wirf, 2005).
In the Northern Territory, J. gossypiifolia has spread into pastoral land, forming dense thickets with little or no grass underneath, and making land unsuitable for grazing (Miller, 1982). In Papua New Guinea, sown pastures have become non-viable due to encroachment of several weeds including J. gossypiifolia (Chandhokar, 1978).
Because of its dense canopy, shallow root system and allelopathic qualities, J. gossypiifolia invasion results in a loss of biodiversity, wildlife habitat, changed fire regimes, increased soil erosion and destabilization of creek and river banks. Monospecific stands of J. gossypiifolia inhibit establishment of native plants (Csurhes, 1999). The impenetrable thickets of dense J. gossypiifolia infestations may also be used as refuges by feral animals such as feral pigs, hindering control of these environmentally damaging pest species (King and Wirf, 2005).
Most parts of the J. gossypiifolia plant contain toxins of various concentrations, posing a serious human health risk. Detrimental effects on human health include seed poisoning (Kingsbury, 1964), dermatitis (Souder, 1963) and sneezing (Irvine, 1961). Whilst numerous cases of severe poisoning have been reported from the plant’s native range, no human deaths were recorded (Begg and Gaskin, 1994).
For Aboriginal people of Australia, J. gossypiifolia can reduce the availability of traditional foods (bush tucker) and other resources by displacing native plants and animals (King and Wirf, 2005). Dense infestations of J. gossypiifolia can also disrupt spiritual and physical connections to the country, for example by restricting access to sacred sites (Gardner 2005, King and Wirf, 2005). Impenetrable infestations of J. gossypiifolia restrict hunting, camping and bushwalking activities and the general physical movement of people (Gardner, 2005; King and Wirf, 2005). J. gossypiifolia is unpleasant to touch because it is sticky and leaves permanent reddish-brown stains on garments. Furthermore, J. gossypiifolia infestations appear dramatically different to the surrounding natural environment and are highly noticeable, diminishing the aesthetic values of the natural landscape (King and Wirf, 2005).
The major benefits of J. gossypiifolia are associated with its medicinal attributes. Various parts have been studied as sources of novel pharmaceuticals, including potential anticancer drugs (Biehl and Hecker, 1985; de Padua et al., 1999). Roots, stems, leaves, seeds, and fruits have been widely used in traditional folk medicine in many parts of western Africa (de Padua et al., 1999; IPCS INCHEM, 2004). Extracts from the plant have been used to treat a number of human ailments, ranging from anaemia, vertigo, worms, leprosy, leukaemia, dysphonia, urinary complaints, ulcers, itches, conjunctivitis, dermatitis, gout, snakebite and venereal diseases (Irvine, 1961; Kupchan et al., 1976; Morton, 1981; Liogier, 1990; Das and Das, 1994; Horsten et al., 1996; de Padua et al., 1999).
In Peru the leaves and latex are used to treat abscesses, tonsillitis, asthma, diarrhoea, toothache, fever, gingivitis, fungal skin infections, inflammations, burns and coughs (Pinedo et al., 1997).
In certain African countries, people are accustomed to chewing seeds of J. gossypiifolia when in need of a laxative (IPCS INCHEM, 2004). The seeds are oily, purgative and emetic (Irvine, 1961). Tea made from bark is used in Nigeria to cure intestinal worms (Irvine, 1961). The leaves are boiled up and used as a bath for fever and the leaves are used as a purgative in Jamaica (Irvine, 1961; de Padua et al., 1999). Roots of J. gossypiifolia have been used for treatment of leprosy (Das and Das, 1994; Baxter, 2000).
Plant parts used for healthcare in India include the young stem, root, bark and latex (Das and Das, 1994). These parts are used either alone or with other components for the treatment of abdominal discomfort, bone fracture, toothache, conjunctivitis, open wounds, diarrhoea, dysentery, haemorrhoids, intra-uterine death, muscular pain, rheumatism, tongue sores and infections around fingernails and toenails (Banerji et al., 1993; de Padua et al., 1999). Crude hot water extract of J. gossypiifolia exhibited anti-malarial properties. It was capable of 100% inhibition of the malaria agent Plasmodium falciparum (Gbeassor et al., 1989).
Extracts of J. gossypiifolia have a reputation as a cancer cure (Biswanth and Ratna, 1995; Biswanth et al., 1996; Morton, 1981, 1982; Taylor et al., 1983). For example, on the island of Aruba, people believe that a decoction of the stems from J. gossypiifolia cures throat cancer (Morton, 1982). Derivatives of the diterpene jatrophone were also isolated from roots of J. gossypiifolia and shown to have anti-tumour properties in vitro (Taylor et al., 1983).
Other potential benefits of J. gossypiifolia include its use as a source of oil for energy (Forni-Martins and Cruz, 1985; Burkill, 1994), a source of plant food for human and animal consumption, an additive for plastic formulations (Ogbobe and Akano, 1993) and a source of insecticides (Prasad et al., 1993; Chatterjee et al., 1980). In Asia, J. gossypiifolia is used for dye (Smith, 1995) and for production of biogas when used as pressed mud cake (Abubacker et al., 1999). Oil extracted from J. gossypiifolia seeds is also used as an illuminant in Africa (Burkill, 1994).
In drier regions of West Africa, J. gossypiifolia is used as a hedge around villages to protect them against bush fires (Irvine, 1961; Ogbobe and Akano, 1993). Some West Africans also believe that J. gossypiifolia has magical powers that protect against snakes, lightning, and violence (Burkill, 1994).
The closest relatives of J. gossypiifolia from karyotype studies were J. multifida and J. curcas, which were also noted as very similar morphologically (Soontornchainaksaeng and Jenjittikul, 2003). J. gossypiifolia confusion with castor oil or the castor bean Ricinus communis may occur in the field as both species may be found in the same habitat. The two species are easily distinguished by the shape of their leaves and fruits. Castor oil or castor bean has considerably larger leaves which grow 15-30 cm across (sometimes larger) with 7 to 9 pointed lobes with serrated edges compared with 3-5 deeply divided lobes with sticky denticulate edges covered with extra-floral nectaries of J. gossypiifolia. The fruit of castor oil or castor bean are also considerably larger (2.5 cm across) and spinescent compared with the glabrous or pubescent J. gossypiifolia pods which may grow to 1.2 cm long and 1 cm wide.
Individual J. gossypiifolia plants are easy to kill with conventional control techniques such as herbicides, machinery and fire (Bebawi and Campbell 2002b, c, d; Bebawi et al.,2004, 2007a; Guterres et al., 2008). However, large-scale seedling recruitment generally occurs following the initiation of control activities. This highlights the need for a long term management strategy to treat seedlings that may emerge while ever a viable seed bank remains. A key priority should be the establishment of a healthy pasture that will compete with J. gossypiifolia seedlings and reduce the opportunities for establishment (Ashley, 1995; Csurhes, 1999).
Chemical Control
Herbicides can cause high mortality of J. gossypiifolia when applied using hand-help equipment to treat scattered to medium infestations (Chadhokar, 1978; Vitelli et al., 1988; Pitt and Miller, 1991). Plants within dense J. gossypiifolia infestations have also been successfully treated by aerial application of herbicides (Vitelli and Madigan, 2002). However, irrespective of the application technique, considerable recruitment often occurs subsequently (Vitelli and Madigan, 2002).
In Papua New Guinea, foliar application of 2,4-D or 2,4,5-T (5–10 g L-1) in water was not effective against J. gossypiifolia (Chadhokar, 1978). Field studies undertaken in north Queensland showed that the addition of a 0.2% v/v wetting agent (poly dimethyl siloxane) increased the activity of 2,4-D acid (10 g L-1), 2,4-D amine (5 g L-1), and 2,4-D ethyl ester (5 g L-1), resulting in mortality rates of 90, 100 and 90%, respectively (Vitelli et al., 1988). In the same trial, amitrole (4 g L-1), 2,4-DP (24 g L-1), 2,4-D/picloram (4.8/1.2 g L-1), fluroxypyr (4 g L-1), glyphosate (2.8 g L-1), metsulfuron methyl (0.12 g L-1) and triclopyr/picloram (4.5/1.5 g L-1) killed 50, 73, 97, 100, 100, 100 and 100% of the treated plants, respectively (Vitelli et al., 1988). Similar results occurred for metsulfuron methyl (98% mortality) in trials in the Northern Territory (Australia), but fluroxypyr (0.3, 0.6 and 0.9 kg ha-1) and glyphosate (0.45, 0.9 and 1.35 kg ha-1) controlled less than 10% of the treated plants (Pitt and Miller, 1991). The addition of wetting agents will significantly increase the effectiveness of herbicides when applied as foliar sprays (Pitt and Miller, 1991; Csurhes, 1999).
In north Queensland, the aerial application of herbicides was tested as a potential method for treating large areas of dense infestations of J. gossypiifolia growing in non-timbered areas (Vitelli and Madigan, 2002). Using a carrier spray volume of 200 litres per hectare, five foliar herbicides (triclopyr/picloram at 450/150 g ha-1, glyphosate at 2.16 kg ha-1, fluroxypyr at 400 g ha-1, metsulfuron methyl at 72 g ha-1, and metsulfuron methyl plus glyphosate at 72 + 867 g ha-1) killed between 92 to 100% of J. gossypiifolia plants (Vitelli and Madigan, 2002). 2,4-D ester at 4 kg ha-1 performed poorly, killing only 63% (Vitelli and Madigan, 2002). Efficacy dropped from 98 to 42% when J. gossypiifolia growing in timbered country was aerially treated using triclopyr plus picloram (Vitelli and Madigan, 2002). In trials undertaken by the Northern Territory government, three rates of metsulfuron methyl (50, 75 and 100 g ha-1) and three rates of glyphosate mixed with simazine (1.5 + 2.0, 1.5 + 4.0 and 3.0 + 4.0 kg ha-1) were aerially applied with carrier spray volumes of 60 L ha-1 to dense stands of mature J. gossypiifolia. Twelve months after application, there were no visible effects from herbicide application (Pitt and Miller, 1991).
Cut stump and basal bark techniques can produce good kills of J. gossypiifolia. In Papua New Guinea, a complete kill was obtained when 2,4,5-T (5–10 g L-1) in diesel was applied to stumps cut at ground level (Chadhokar, 1978). Similarly, cut stump applications of picloram/2,4,5-T (2.5/10 g L-1) in water and basal bark applications of picloram/2,4,5-T (2/8 g L-1) in diesel killed 100% of plants (Pitt and Miller, 1991). Cut stump trials in north Queensland also found 2,4-D (1.5, 3 and 30 g L-1 water), 2,4-D/picloram (0.5/2, 1/4, and 2/8 g L-1 diesel), fluroxypyr (3, 6 and 12 g L-1 diesel), triclopyr (4.8, 9.6 and 19.2 g L-1 diesel) and neat diesel killed 90 to 100% of the treated J. gossypiifolia plants (Vitelli et al., 1988). To be effective the cut stump method requires both treatment of the stump and disposal of the cut plant, as cut sections can produce seed.
The use of residual herbicides to reduce seedling recruitment has been tested. In the Northern Territory (Australia), high mortality of mature plants and control of seedlings for up to two years was achieved with hexazinone and tebuthiuron (Pitt and Miller, 1991). In a north Queensland trial, metsulfuron methyl applied between 36 and 1162 g ha-1 and tebuthiuron (500, 1000 and 1500 g ha-1) were applied to pots containing 50 J. gossypiifolia seeds. One hundred and eighty days post application, seedling mortality plus seed mortality was recorded at greater than 80% for each treatment, compared to control pots with 35% mortality (JS Vitelli, Biosecurity Queensland, Australia, personal communication, 2009).
Fire
Burning is an effective control technique against J. gossypiifolia where there is sufficient fuel to carry a fire (Bebawi and Campbell, 2002a, b, c; Guterres et al., 2008). In a field trial in northern Australia, 76% mortality occurred following burning of a riparian infestation. A second burn a year later increased the mortality rate to 92% (Bebawi and Campbell, 2002c), with temperatures of up to 640°C recorded at ground level (Bebawi and Campbell, 2002c). Exposure of J. gossypiifolia to such high temperatures caused them to ooze caramelized latex and blister profusely due to the high sugar concentration of the latex (Bebawi and Campbell, 2002b). Despite the average fuel load being relatively high, there was significant variation across the site and mortality tended to vary accordingly. Nevertheless, juvenile plants were more susceptible to fire than mature plants, with old plants being the most tolerant (Bebawi and Campbell, 2002c).
A large portion of the J. gossypiifolia seed bank is able to survive fire, resulting in large scale recruitment afterwards (Bebawi and Campbell, 2002c). For example, in the fire trial mentioned previously, 540 seedlings m-2 emerged from burnt plots compared with 190 seedlings m-2 in unburnt plots. Seedling density in burnt plots averaged 37 seedlings m-2 after two years, compared with four seedlings m-2 in the unburnt controls (Bebawi and Campbell, 2002c). Fire could therefore exacerbate the J. gossypiifolia problem if follow-up control is not implemented.
Fire can only be used when there is a sufficient fuel load (Vitelli, 2000). In drier areas, burns may only occur following several years of above average rainfall. There are also substantial costs in using fire, particularly as stock must be excluded before and after a burn, fire cannot always be used in fire sensitive areas (Vitelli, 2000) and pasture may be lost.
Flame throwers have been tested for treatment of J. gossypiifolia where chemical and mechanical control is inappropriate or ineffective (Vitelli and Madigan, 2004). Flaming for 10 seconds around the entire circumference of the base of individual J. gossypiifolia plants (5 cm above ground level) at a maximum temperature of 820°C killed 92% of the treated plants at a cost of 7.5 cents per plant (Vitelli and Madigan, 2004). This practice would be most appropriate for treating small and/or scattered infestations.
Mechanical Control
Since J. gossypiifolia has a shallow root system it can be fairly easily removed by hand, especially if the soil is moist (Csurhes, 1999). In Australia, a number of landholders with small patches prefer to use this practice. By physically removing the plants they are confident that plants have been killed. Landholders also often put the removed material in stacks and burn it, so as to kill any reproductive material and reduce the risk of plants re-attaching to the soil and growing (SD Campbell, Tropical Weeds Research Centre, Australia, personal communication, 2009).
Cutting J. gossypiifolia close to ground level can also be effective (Bebawi and Campbell, 2002b; Guterres et al., 2008). In a cutting trial, no plants cut off at ground level regrew, whereas those cut above ground level reacted differently depending on the season. The majority of plants cut at 10, 20, and 40 cm height in the dry season regrew, whereas only those cut at 20 and 40 cm in the wet season regrew. In general, plants were more susceptible when cut in the wet season (Bebawi and Campbell, 2002b). Similar results were recently reported by Guterres et al. (2008).
The use of tractor-mounted slashers has provided similar results in areas of suitable terrain. In one field trial 100% of plants were killed, irrespective of whether they were juvenile, mature or adult plants (Bebawi et al., 2004). There are, however, instances in other areas where not all plants have been killed after being cut off at ground level, with re-shooting sometimes occurring (Chadhokar, 1978; Pitt and Miller, 1991).
The use of heavy machinery such as bulldozers can be effective for directly killing plants, as well as for cleaning up infested areas so that other forms of control can be implemented. In one trial, a stick rake on the front of a bulldozer caused greater than 90% mortality of plants within a dense infestation. Those that survived were generally smaller plants that escaped being ripped out (Bebawi et al., 2004). For mechanical control of woody weeds such as J. gossypiifolia, efficacy is generally highest for treatments that completely remove the whole plant or sever the root system underground (20–30 cm). Season of application, soil texture, soil moisture and weed density also influence survival and regrowth (Vitelli, 2000).
Physical disturbance that occurs when using mechanical control will generally result in massive recruitment of seedlings (Bebawi et al., 2004). This can be a means of depleting the seed bank more quickly, provided regrowth is controlled before it reaches reproductive maturity (Campbell and Grice, 2000).
Pasture Management
Preliminary results from a field trial in north Queensland investigating the impact of five simulated grazing regimes on four J. gossypiifolia population densities suggest that J. gossypiifolia grows best in areas void of pasture. Where there is grass cover, seedling recruitment is reduced and plants grow more slowly (Bebawi et al., 2006, 2007c). Over four years, buffel grass-dominated (Cenchrus ciliaris) plots reduced flowering of J. gossypiifolia plants by 89% and seed production incidence by 95% compared with plots void of pasture (Bebawi et al., 2007c). Furthermore, after seven years, minimal mortality (4%) has occurred in areas devoid of pasture. In contrast, an average of 53% and 68% mortality occurred after four and seven years in pastured plots, respectively (Bebawi et al., 2006; DEEDI, 2009). These findings suggest that J. gossypiifolia is likely to dominate areas faster if they are void of pasture (Bebawi et al., 2006) and that maintenance of a competitive pasture will help greatly in the management of this weed. Nevertheless, as for most tropical weeds, there are very few data on how J. gossypiifolia can be managed using grazing or pasture management.
Biological control
Jatropha spp. generally have few phytophagous insects or pathogens in their native range (Dehgan, 1982). In India, J. gossypiifolia was reported completely free of any visible fungal or insect damage (Raina and Gaikwad, 1987). However, in Australia, the leaf-mining moth, Epicephala sp. (Wilson, 1997) and the castor oil looper, Achaea janata cause minor defoliation of J. gossypiifolia. The tenebrionids beetles, Lyphia australis and Platycotylus nitidulus, and the nitidulid beetles, Carpophilus marginellus and C. obsoletus, have also been observed attacking the stems in the Northern Territory (FF Bebawi, Biosecurity Queensland, Australia, personal communication, 2009).
A soil borne fungus Scytalidium dimidiatum has been observed to cause canker and wilt of J. gossypiifolia stems in north Queensland. S. dimidiatum (formerly Torula dimidiata) is the synanamorph of Nattrassia mangiferae (syn. Hendersonula toruloidea). It is a very common soil-borne fungus associated with a wide range of hosts, including crop species (such as mangoes in Queensland) and eucalypts, and causing stem end rot (Tomley, 2003).
J. gossypiifolia has been a target for biological control in Australia since 1996 (Heard and Chan, 2002; Heard et al., 2009). A total of 170 locations in nine countries (Mexico, Venezuela, Dominican Republic, Puerto Rico, Nicaragua, Netherlands Antilles, Guatemala, Trinidad and Cuba) have been investigated for potential agents against Jatropha. Over 1000 specimens were collected and sent for identification.
Of the insects investigated, Pachycoris klugii (Scutelleridae) was the most damaging herbivore of plantations of J. curcas in Nicaragua, but unfortunately failed to develop on J. gossypiifolia. Cylindrocopturus jatrophae (Curculionidae), a stem-boring weevil, Colaspis sp. (Chrysomelidae) and Styloleptus sp. and Parmenonta sp. (Cerambycidae) were imported into quarantine but could not be reared. While Lagocheirus sp. (Cerambycidae) established in quarantine, host testing revealed it fed on cassava and was rejected.
To date, the seed feeding jewel bug, Agonosoma trilineatum (Scutelleridae) is the only agent approved for release in Australia against J. gossypiifolia. The jewel bug was first released in the Northern Territory in March 2003 and in north Queensland in June 2003 (Heard, 2003; Bebawi, 2004; Heard et al., 2009).
The potential impact of jewel bugs on J. gossypiifolia has been quantified in a greenhouse experiment (Bebawi et al., 2005a, 2007d). Potted J. gossypiifolia plants were exposed to 0, 6 or 24 jewel bugs per plant. Jewel bugs significantly increased the level of abortion of both immature and mature capsules, particularly at the higher density of insects. Immature capsules were the most susceptible, averaging 80% capsule abortion, compared with only 21% for more mature capsules (Bebawi et al., 2005a, 2007d). After 30 days exposure to the jewel bug, seed production was reduced by 22% and 75% respectively at low and high jewel bug density (Bebawi et al., 2005a, 2007d). Furthermore, nearly 60% and 83% of the seeds produced at low- and high jewel bug density were damaged (Bebawi et al., 2005a, 2007d). In the laboratory, feeding by adult jewel bugs completely destroyed seeds of J. gossypiifolia (Heard and Chan, 2002; Heard et al., 2009).
To date, there have been limited signs of A. trilineatum establishment at five of the fifteen release sites in Queensland. The most promising signs of establishment were seen in 2007, when both adults and nymphs were observed at two sites following release. Evidence of feeding damage on seed capsules was also noted. It is uncertain yet whether the insect will persist at these sites (CJ Lockett, Tropical Weeds Research Centre, Australia, personal communication, 2009).
In 2007 the J. gossypiifolia biocontrol program was recommenced in Australia to keep looking for other potentially damaging insects or pathogens in Central and South America (Randall et al. 2009). Currently, only two agents are available for further testing – a rust fungus (Phakopsora jatrophicola) and a stem-boring weevil (Cylindrocopturus imbricatus).
Integrated Management
No single control method provides effective management of J. gossypiifolia at a reasonable cost (Vitelli, 2000). Results from a trial of integrated control methods have shown that high kill rates of J. gossypiifolia can be obtained with a single application of techniques such as foliar spraying, slashing, fire and stickraking (Bebawi et al., 2004). However, follow-up control will generally be needed for several years to control remaining plants and seedling regrowth. Selective foliar spraying appears to be one of the most effective options for follow-up as it allows the maintenance of a grass cover to compete with seedlings that emerge afterwards. In the integrated control trial, for every plant killed by a single foliar spray treatment, 20 plants were recruited from the seed bank, much less than in slashing, mechanical and fire treatments which averaged 97, 74 and 69 plants, respectively (Bebawi et al., 2004). Furthermore, of all the treatment combinations implemented, repeated foliar spraying on an annual basis was the only one to reduce the density of bellyache bush to zero. This was achieved after two follow-ups following the initial primary application. Foliar spraying as a follow-up to slashing, stick-raking and burning also proved effective (average 99% reduction) but did not remove all bellyache bush plants within the study period. Slashing as a follow-up caused an average population reduction of 97% after being undertaken threes times (on an annual basis), irrespective of the initial primary treatment. Similarly, burning following foliar spraying resulted in a 97% decline in the bellyache bush population (FF Bebawi, Biosecurity Queensland, Australia, personal communication, 2009).
In Australia, research into control strategies has identified techniques that can be integrated to effectively treat infestations growing in pasture situations. Control of J. gossypiifolia in more sensitive environments, such as within watercourses is more problematic and warrants investigation.
A recent molecular genetic study identified that multiple introductions of diverse haplotypes from throughout the native range has occurred in Australia and may explain some of the considerable variation that is found between infestations (Prentis et al., 2008). Several biotypes have also been noted based on morphological, phenological, and physiological differences (Pitt and Miller, 1991; Bebawi and Campbell, 2004; Bebawi et al., 2007e, 2009). Detailed taxonomic, genetic and ecological studies are now required to verify the differences which could have implications for management, particularly selection of biological control agents. Such studies would also help determine whether there is only one variety present or if some of the noted biotypes could in fact be different varieties. It has been suggested that two of the biotypes in Australia could potentially be J. gossypiifolia var. elegans and J. gossypiifolia var. staphysagrifolia (B Dehgan, University of Florida, USA, personal communication, 2008).
Hnatiuk RJ, 1990. Census of Australian Vascular Plants. Australian Flora and Fauna Series Number 11. Canberra, Australia: Australian Government Publishing Service.
Holm LG; Pancho JV; Herberger JP; Plucknett DL, 1979. A Geographical Atlas of World Weeds. New York, USA: Wiley.
Pitt JL, 1999. Bellyache Bush (Jatropha gossypifolia). Agnote No. 480. Darwin, Australia: Department of Primary Industries and Fisheries, Northern Territory of Australia.
