Chrysanthemoides monilifera (boneseed)

Summary of Invasiveness

Top of pageThe success of C. monilifera in Australia is largely due to its vigorous growth and ability to numerically swamp the seed bank with viable seed (Noble and Weiss, 1989; Parsons and Cuthbertson, 1992). This is aided by an absence of natural enemies and its ability to regenerate quickly after fire. Studies have shown that, after burning, C. monilifera responds much more than native species to the increased soil nutrient levels, particularly phosphorus (Parsons and Cuthbertson, 1992). With the large seed bank in the soil there is dense growth of seedlings after fires. Furthermore, C. monilifera has a more efficient leaf arrangement than most native plants, enabling it to shade smaller dune species effectively. It also develops a more vigorous root system than some of its competitors, which is an advantage during periods of moisture stress (Parsons and Cuthbertson, 1992).

Subsp. rotundata has a more extensive root system than subsp. monilifera and appears to be more aggressive and difficult to control (Parsons and Cuthbertson, 1992).

Notes on Taxonomy and Nomenclature

Top of pageThe genus Chrysanthemoides has two species, Chrysanthemoides incana (Burm. F.) T. Norl. and Chrysanthemoides monilifera (L.) T. Norl. Norlindh (1943) split C. monilifera into six subspecies: C. monilifera subsp. canescens (DC.) T. Norl., C. monilifera subsp. monilifera T. Norl., C. monilifera subsp. pisifera (L.) T. Norl., C. monilifera subsp. rotundata (DC.) T. Norl., C. monilifera subsp. septentrionalis T. Norl. and C. monilifera subsp. subcanescens (DC.) T. Norl. (Scott, 1996).

Chrysanthemoides means 'chrysanthemum-like,' chrysanthemum meaning golden flower from the Greek words chryos (= gold) and anthos (= flower), hence it describes the flower colour (Parsons and Cuthbertson, 1992). Monilifera is from the Latin word monile, meaning 'necklace,' because of the bead-like fruit formed as a ring in the flower head.

Description

Top of pageC. monilifera is a perennial shrub, between 1 and 3 m high. Roots are shallow, with no distinct taproot. Leaves are alternate, 3-8 cm long, ovate, tapering at the base, irregularly serrated or toothed, shortly stalked, practically hairless except for a cottony down on young leaves. Inner phyllaries are ovate or lanceolate. Stems are woody, branched and upper stems are often purplish. Florets are bright yellow, in shortly stalked heads, 2-3 cm in diameter, clustered at the ends of branches; ray florets (petals) 5-13 per head. Fruits have green fleshy skin at first, becoming black then flaking off to leave a hard, whitish inner coat, 5-8 mm in diameter. Seeds are very hard and bone-like in colour and texture when ripe; there is a single seed in each fruit.

The main differences between subsp. rotundata and subsp. monilifera are: rotundata is a sprawling shrub, whereas monilifera is erect; the inner phyllaries of rotundata are narrowly ovate or lanceolate, whereas monilifera is broadly ovate or lanceolate; rotundata commonly has between 11 and 13 ray florets per head, whereas monilifera has 5-6 ray florets per head; the seeds of rotundata are longitudinally ribbed, whereas those of monilifera are more or less smooth; and the roots of rotundata can be produced on prostate stems that come in contact with the soil.

Plant Type

Top of pagePerennial
Seed propagated
Shrub
Woody

Distribution

Top of pageC. monilifera is native to the sandy soils of southern and south-eastern South Africa, where there are six subspecies (Stuart, 2002). In South Africa, subsp. rotundata is only moderately abundant and individual plants have been described as 'small and sparse' in appearance.

History of Introduction and Spread

Top of pageC. monilifera has been both deliberately and accidentally introduced into other countries. It is known to have been introduced as a garden ornamental. For example, subsp. monilifera was introduced into Australia as an ornamental plant in the 1850s and was grown in gardens in Sydney, Melbourne and Adelaide (Parsons and Cuthbertson, 1992). In New Zealand, subsp. monilifera was also introduced as an ornamental plant and was first recorded growing wild in 1870 (Mallinson, 2003). Further spread is expected to have occurred via the dumping of garden refuse into nearby bushland and the distribution of its fleshy fruits by possums, foxes and birds. Rabbits were known to feed on the seedlings of this subspecies in Australia, and the reduction of rabbit numbers in the 1950s is expected to have contributed to the weed's rapid spread throughout coastal bushland at that time (Parsons and Cuthbertson, 1992).

C. monilifera has also been accidentally introduced into other countries. For example, subsp. rotundata appears to have been introduced into Australia via the dumping of ship's ballast on the banks of the Hunter River around 1908 (Stuart, 2002). Later recognized as an effective colonizer on sandy soils, it was planted for this purpose between 1946 and 1968 along much of the New South Wales coastline in areas that had been sandmined or otherwise disturbed (Barr, 1965; Parsons and Cuthbertson, 1992). Not only did it colonize these areas, it soon dominated the vegetation. Further spread via the dispersal of its seed has been facilitated by birds, foxes and sand movement.

Risk of Introduction

Top of pageFurther spread of C. monilifera may occur if its popularity as an ornamental garden plant increases. However, it is not a contaminant of crop seed or other agricultural produce, and currently there is more awareness with regard to the disposal of ship ballast compared to the early 1900s.

C. monilifera is prohibited as a noxious weed in Australia, being rated as a weed of national significance (Kriticos and Groves, 2000). Subsp. monilifera is prohibited from sale or propagation in New Zealand (Mallinson, 2003).

Habitat

Top of pageC. monilifera is a common weed of subtropical and subhumid scrublands (Parsons and Cuthbertson, 1992). It prefers sandy or medium-textured soils and disturbed situations, particularly near the sea where it tolerates saline conditions.

In Australia, subsp. rotundata grows in a range of environments from open, exposed dunes to shaded forests. As a result, in Australia, it is known to have invaded sand dune heathlands and grasslands, headland heathlands and grasslands, coastal woodlands, coastal dry schlerophyll forests, and littoral rainforests. Mallee vegetation in western New South Wales has also been affected by this subspecies.

The lower sprawling growth of subsp. rotundata equips it well to survive on windy coastal sites (Parsons and Cuthbertson, 1992). The more erect growth of subsp. monilifera makes it better suited to competing with taller forest plants. Subsp. monilifera commonly inhabits rocky, infertile and inhospitable sites where few other plant species are able to establish.

C. monilifera has a variable climate preference, with subsp. monilifera commonly found in hot summer/cold winter latitudes, whereas subsp. rotundata occurs in both hot summer/warm winter and hot summer/cold winter latitudes.

Hosts/Species Affected

Top of pageC. monilifera is not normally a weed of crops. In coastal areas, subsp. rotundata commonly displaces native Australian coastal species such as Acacia longifolia, Lomandra longifolia, Correa alba, Leucopogon parviflorus and Metrosideros excelsa (in New Zealand) to form almost pure stands in many areas (Parsons and Cuthbertson, 1992; Mallinson, 2003). No plant species is known to have become extinct as a result of the subsp. rotundata invasion in Australia, but its distribution does overlap with those of some rare and endangered plant species, notably Pimelea spicata, Zieria prostrata, Cynanchum elegans and Thesium australe.

Growth Stages

Top of pageFlowering stage, Fruiting stage, Pre-emergence, Seedling stage, Vegetative growing stage

Biology and Ecology

Top of pagePhysiology and Phenology

Seeds of C. monilifera germinate at any time of the year but mostly in autumn. Research conducted on the germination of subsp. rotundata seed found that the germination process is closely related to the permeability of the seed coat and/or associated structures (Aveyard, 1971). Water uptake by the seed during the first 98 hours is very low. However, the uptake provides the necessary force to crack the seed coat, thereby permitting the entry of free water and a fruition of the germination process. It was also shown that the seed coat, and not the surrounding aril (part of the seed), was responsible for restricting the moisture uptake of a seed. The optimal germination temperature for subsp. rotundata seed is 25°C.

C. monilifera growth is rapid during the winter and a few plants may flower in the first year, particularly on burnt areas where there is little competition (Parsons and Cuthbertson, 1992). Bending of shoots, such that distal sections are relatively higher, is the common growth habit of subsp. rotundata (Stuart, 2002). Typically, as the shoots extend, they lean and may become prostate on the ground or become supported on other lower shoots and branches (Stuart, 2002). The tendency to lean is exacerbated under dense shade where the shoots tend to be etiolated (Stuart, 2002). These horizontal stems typically give rise to vertical axillary branches, until these branches lean in turn (Stuart, 2002). This branching dynamic creates a complex morphology in which the uppermost shoots are most likely to flower, but as they bend and lean they are likely to be overtopped and thus, less likely to flower. As a result, a shoot may be in a prime site for flowering in one season, but be overtopped by the next.

It has been suggested that flowering may be initiated by a threshold level of shoot growth (i.e. resource accumulation) (Stuart, 2002). However, more research is needed to validate this hypothesis.

Mostly, C. monilifera plants are at least 18 months and sometimes 3 years old before flowering (Parsons and Cuthbertson, 1992). Once a vegetative shoot tip becomes switched to inflorescence production, flowers develop at the apex, and there is an associated suppression of further local vegetative shoot growth until flower and fruit maturation ceases (Stuart, 2002). Axillary buds are then able to commence vegetative growth, typically resulting in shoots developing from the proximate 1-5 leaf axils. There is considerable synchrony in shoot initiation of flowering in a plant. However, shoots of subsp. rotundata have been observed to not initiate flowering all at the same time, and slow-growing shoots are less likely to flower.

In its native range of South Africa, flower production by C. monilifera is highly seasonal (Scott, 1996). In the Mediterranean-type climate of the Cape Peninsula, subsp. monilifera and subsp. pisifera flower in spring. Flowers are absent only during the dry period of late summer to autumn. In the southern and eastern Cape Province, where rain commonly falls throughout the year, flowering occurs during spring to early summer and flowers are absent during late summer and autumn. In the subtropical climate of Natal, some flowers are present on subsp. rotundata throughout the entire year. The presence of ripe fruits is highly seasonal in the Cape Province, but in Natal ripe fruits are present throughout the year, or at irregular intervals.

In Australia, flowers of subsp. monilifera are formed in late winter and spring but seeds are not shed until summer (Parsons and Cuthbertson, 1992). Flowers of subsp. rotundata are produced almost year-round, with northern New South Wales populations having two flowering peaks during the year (in autumn and spring), and only one flowering peak in southern New South Wales populations (in autumn) (Stuart, 2002). Parsons and Cuthbertson (1992) state that about 60% of seeds are viable. The hard seed coat splits open in some, and these germinate as soon as soil moisture is adequate, but with many the seed coat remains intact.

On the basis of ring counts of subsp. rotundata, C. monilifera has been shown to live for upwards of 35 years (Stuart et al., 2002).

Reproductive Biology

The flowers of C. monilifera are insect-pollinated. Propagation is completely by seeds, which can be produced in large numbers. Up to 50,000 seed per plant per year have been recorded within populations in Australia (Parsons and Cuthbertson, 1992).

The viability of seed in the soil is reduced after 2 years and only a small number live more than 5 years. However, a few seeds remain dormant for at least 10 years. Viability is retained longer with increased depth of burial.

Environmental Requirements

C. monilifera can successfully grow in coastal foredunes, coastal bushland and partially shaded habitats. It grows well in both sheltered hind-dune zones and less protected foredune zones. It is tolerant of salinity, strong wind, wind-blown sand, salt and water, drought, low nutrients and, to some extent, of disturbances such as fire. Subsp. rotundata grows poorly in wet or swampy soils and has a low tolerance to frost.

Rain Regime

Top of pageBimodal
Summer
Winter

Notes on Natural Enemies

Top of pageIn South Africa, Mesoclanis species have been found to be responsible for up to 9% loss of potential, pre-dispersed seed (Scott, 1996). Other insects, principally unidentified Lepidoptera larvae, commonly destroy 4-25% of potential seed. Rodents are the cause of most identifiable post-dispersal damage. Most of the rodent damage was probably due to the striped mouse (Rhabdomys pumilio), an abundant seed-eating rodent that was often observed. A number of herbivores are also responsible for defoliation. Subsp. rotundata is commonly attacked by the chrysomelid Ageniosa electoralis.

In Australia, C. monilifera supports a depauperate phytophagous fauna, and although leaf-feeding insects have been observed, e.g. Chrysodeixis eriosoma, Teia anartoides, Epiphyas postvittana and Phlyctinus callosus, these insects cause only minor damage (Adair and Scott, 1991).

Means of Movement and Dispersal

Top of pageNatural Dispersal (Non-Biotic)

Propagation is via seed, which are produced in large numbers. Short-distance dispersal can be via wind; however, seeds commonly fall beneath the mother plant. The fruits and seeds are also effectively carried by running water or by sand movement.

Vector Transmission (Biotic)

Birds are a major agent of spread because they readily eat the fleshy fruit, and seed is either passed or regurgitated in a viable condition (Aveyard, 1971; Parsons and Cuthbertson, 1992). Rabbits, foxes (Meek, 1998) and cattle also eat the fruit and spread the seeds in their droppings, whereas ants have been observed carrying the fruit to their nests where they eat the fleshy skins and discard the seeds.

Accidental Introduction

The main incident of accidental introduction has been via contaminated ship ballast from South Africa. Within Australia, contaminated gravel carted from infested areas has been one of the main agents of spread away from the infestation at the You Yangs in Victoria (Parsons and Cuthbertson, 1992).

Intentional Introduction

Intentional introduction of C. monilifera is quite likely, as it has been used in the past as an ornamental for gardens.

Impact

Top of pageNo figures are available for the total cost of C. monilifera infestation. In Australia, the costs of large-scale management to date are expected to be high, as they involve considerable amounts of labour from Dunecare and Landcare groups, aerial spraying with herbicides, bulldozing and establishing a biological control programme.

Environmental Impact

Top of pageC. monilifera does not affect agriculture detrimentally and is rarely found in pastures. However, it is a weed because it is readily able to establish in areas of native vegetation ranging from mallee scrub to wet schlerophyll forests, whether disturbed or not, and can eventually out-compete and eliminate many native species. It establishes on disturbed areas and, hence, is often first associated with roadworks, land clearing and burning. It also establishes on sites of minimum disturbance, but at a much slower rate.

Impact: Biodiversity

Top of pageThe presence of C. monilifera affects wildlife. Parsons and Cuthbertson (1992) note that as a dense stand, C. monilifera destroys, or drastically alters, the habitat of native birds and animals in Australia. Also, it has replaced plant species known to be important food sources for migratory birds. It is also known to replace plants that are deemed important to apiarists as a source of nectar and pollen during winter. However, the importance of C. monilifera itself, as a food plant for bees, has not been determined.

C. monilifera is able to develop rapidly into a dense shrub with high foliage cover, and its shoots lean and compress shorter plants beneath the canopy where they remain light limited (Stuart, 2002). Native understorey species such as Acacia longifolia and Lomandra longifolia are readily displaced by C. monilifera on both sand dunes and in adjacent open and woodland habitats. By shading understorey plants and stopping their seed production, as well as inhibiting further germination from the native seed bank, C. monilifera is able to completely replace native vegetation in an area. After time, the native seed bank is diminished, so recovery is limited even if C. monilifera is removed.

Social Impact

Top of pageC. monilifera commonly forms dense infestations in coastal areas. This often seriously interferes with access to beaches.

Uses

Top of pageC. monilifera is not used for culinary or medicinal purposes in Australia although it is possible to make jam from the fruits.

Prevention and Control

Top of pageCultural Control

C. monilifera is not known to persist when grazed and trampled by stock or when cultivated (Parsons and Cuthbertson, 1992). Fire can also play an important role in the control of C. monilifera. Fire destroys seedlings and many mature plants, and stimulates almost all seed in the soil to germinate. This means that elimination of the weed may be possible, if all the resultant seedlings can be controlled. Mature plants of subsp. rotundata are less susceptible to fire and strong resprouting often occurs after a fire. Adult plants of subsp. monilifera are generally killed by fire, but 26% of adult subsp. rotundata generally regenerate from the base of the mature plant after fire (Weiss, 1984; Scott, 1996).

Mechanical Control

On undeveloped land, small plants can be hand-pulled and larger plants can be grubbed or pulled with a tractor (Parsons and Cuthbertson, 1992). As this treatment is only effective when soil conditions allow the bulk of the root system of each plant to be removed, the shallow root system of C. monilifera makes pulling a practical treatment in contrast to most other woody weeds where deeper root systems prevent such a control measure.

The removal of adult C. monilifera plants generally stimulates seed germination and, naturally, the resulting seedlings must be removed before they produce further seeds.

Chemical Control

There are several herbicides, including bromoxynil, glyphosate and picloram, which kill C. monilifera either by overall spraying or painting onto the cut stumps immediately after removal of the top growth (Parsons and Cuthbertson, 1992). 2,4-D amine is also very effective as a cut-stump treatment, and amitrole and metsulfuron-methyl are used in some areas as overall sprays applied at any stage of growth.

Observations showed that as larger bushes died from herbicide (glyphosate) application, prolific growth of freshly germinated seedlings occurred beneath the desiccated bushes (Cooney et al., 1982). Therefore, for satisfactory long-term control, repeated spraying is warranted until the seed supply is exhausted or surrounding native vegetation suffocates the emerging seedlings.

Aerial application of herbicides is not generally used on C. monilifera because it is non-selective, but experience in Australia using glyphosate indicates that much of the population can be removed in this way with little damage to native species (Parsons and Cuthbertson, 1992).

Biological Control

Noble and Weiss (1989) measured the movement of buried subsp. rotundata seed in the soil. From their research, a mathematical model was developed, which predicted that pre-dispersed seed predation by a biological control agent would have to reduce the potentially germinable seed population by more than 95% to provide satisfactory control.

A number of biological control agents have been released within C. monilifera populations in Australia. The host preferences of the biological control agents vary between subsp. monilifera and rotundata. Of the biological control agents released in Australia, three have established and caused damage: Comostolopsis germana, a tip moth released in 1989 (Adair and Scott, 1991); Mesoclanis polana, a seed-predating fly released in 1996; and Tortrix sp., a leafroller released in 2000 (Kriticos and Groves, 2000). As well as the insects already released, a fungus and an eriophyiid mite are under consideration for release.

An assessment of the three species of Mesoclanis spp. seed flies present in South Africa as biological control agents for the two subspecies of the weed in Australia is provided by Edwards (1998). An evaluation of the likely impact of the rust fungus Endophyllum osteospermi as a biological control agent is provided by Wood (2002).

Stuart et al. (2002) modelled the population dynamics of subsp. rotundata and the biological control agent M. polana. The results of this research emphasised that effective biological control of C. monilifera may require multiple agents that each influence different aspects of an adult plant's life cycle.

Integrated Control

Seedlings of C. monilifera are susceptible to bromoxynil and an effective control programme involves burning infested areas when conditions are suitable, waiting for the seeds to germinate, and then spraying with bromoxynil (Parsons and Cuthbertson, 1992). Native vegetation is seldom destroyed by fire and most native species tolerate bromoxynil, hence, C. monilifera can be effectively and selectively removed by this technique.

For large and extensive infestations of C. monilifera in coastal heath, woodlands and grasslands on hind dunes, a multi-stage spray-burn-spray strategy incorporating biological control agents is recommended. Herbicide spraying should be conducted in winter and some areas should be left unsprayed to allow biological control agents to persist and subsequently disperse. However, fire should not be used on the foredunes where the ecosystem is much more sensitive to erosion.

References

Top of page

Adair RJ, Scott JK, 1989. The life-history and host specificity of Comostolopsis germana Prout (Lepidoptera: Geometridae), a biological control agent of Chrysanthemoides monilifera (Compositae). Bulletin of Entomological Research, 79(4):649-657

Adair RJ, Scott JK, 1991. Distribution, life history, host specificity and suitability of an undescribed Chrysolina species (Coleoptera: Chrysomelidae) for the biological control of Chrysanthemoides monilifera (Compositae). Bulletin of Entomological Research, 81(3):235-242

Aveyard JM, 1971. Studies on the germination of Bitou Bush, (Chrysanthemoides monilifera (L) T. Norl). Journal of the Soil Conservation Service of New South Wales, 27:82-91.

Barr DA, 1965. Restoration of coastal dunes after beach mining. Journal of the Soil Conservation Service of NSW, 21:199-209.

Cooney PA, Gibbs DG, Golinski KD, 1982. Evaluation of the herbicide 'Roundup' for control of bitou bush (Chrysanthemoides monilifera). Journal of the Soil Conservation Service of N.S.W., 38(1):6-12

Edwards PB, 1998. Seasonal abundance and parasitism of Mesoclanis seed flies (Diptera: Tephritidae) in South Africa, and implications for the biological control of Chrysanthemoides monilifera (Asteraceae) in Australia. Bulletin of Entomological Research, 88(4):407-414; 9 ref.

Kriticos D, Groves R, 2000. Report of the 2nd Bitou Bush Modeling Workshop. CRC for Australian Weed Management.

Mallinson R, 2003. Plant Pest Control Boneseed Chrysanthemoides monilifera Fact Sheet PP08/98. Whakatane, New Zealand: Environment B.O.P.

Meek PD, 1998. 'Weed seeds and whoopsie daisies': viability of bitou bush Chrysanthemoides monilifera seeds in fox (Vulpes vulpes) scats. Plant Protection Quarterly, 13(1):21-26; 20 ref.

Noble IR, Weiss PW, 1989. Movement and modeling of buried seed of the invasive perennial Chrysanthemoides monilifera in coastal dunes and biological control. Australian Journal of Ecology, 14:55-64.

Norlindh T, 1943. Studies in the Calendulaea I. Monograph of the Genera Dimorphotheca, Castalis, Osteospermum, Gibbaria and Chrysanthemoides. CWK. Gleerup, Lund.

Parsons WT, Cuthbertson EG, 1992. Noxious Weeds of Australia. Melbourne, Australia: Inkata Press, 692 pp.

Scott JK, 1996. Population ecology of Chrysanthemoides monilifera in South Africa: implications for its control in Australia. Journal of Applied Ecology, 33(6):1496-1508; 39 ref.

Stuart RM, 2002. A biological control model for Chrysanthemoides monilifera ssp. rotundata (Asteraceae) by Mesoclanis polana (Tephritidae). Honours thesis, Australian National University, Canberra.

Stuart RM, Kriticos DJ, Ash JE, 2002. Modelling the biological control of bitou bush (Chrysanthemoides monilifera: Asteraceae) by Mesoclanis polana (Tephritidae). 13th Australian Weeds Conference: weeds "threats now and forever?", Sheraton Perth Hotel, Perth, Western Australia, 8-13 September 2002: papers and proceedings, 591-594; 7 ref.

Weiss PW, 1984. Seed characteristics and regeneration of some species in invaded coastal communities. Australian Journal of Ecology, 9:99-106.

Wood AR, 2002. Infection of Chrysanthemoides monilifera ssp. monilifera by the rust fungus Endophyllum osteospermi is associated with a reduction in vegetative growth and reproduction. Australasian Plant Pathology, 31(4):409-415; 23 ref.

Distribution Maps

Top of page