Reviews the Previously Separate Species C

Home , Brasenia, Cabombaceae, Vallisneria spiralis

154 Plant Protection Quarterly Vol.12(4) 1997 and C. palaeformis Fassett (Ørgaard 1991). Only one species, Cabomba caroliniana, is known to be naturalized in Australia. The current definition of this species includes Reviews the previously separate species C. australis and C. pulcherrima, and several natural and horticultural varieties (Ørgaard 1991). Three types of cabomba (C. caroliniana var. caroliniana, C. caroliniana var. multipartita and C. caroliniana var. pulcherrima) from The Biology of Australian Weeds eight sites in Florida assessed for genetic 32. Cabomba caroliniana Gray diversity were found to be genetically in- distinguishable. The differences were ecophenotypic rather than genotypic A.P. MackeyA and J.T. SwarbrickB A (Wain et al. 1983, 1985). Ecophenotypic Land Protection, Department of Natural Resources, Locked Bag 40, plasticity is well known in aquatic plants Coorparoo DC, Queensland 4151, Australia. (Sculthorpe 1967). Godfrey and Wooten B Weed Science Consultancy, 15 Katoomba Crescent, Toowoomba, (1981) and Martin and Wain (1991) report Queensland 4350, Australia. that cabomba with high levels of purple pigment grows in very warm waters but that plants from cooler waters have little Name purple pigment and are green. Con- Cabomba is an aboriginal American word dissected leaves, long-stalked hypo- versely, Leslie (1985) reports aquarists as for aquatic plant. The specific name gynous flowers with three sepals and saying that the purple colour develops in caroliniana indicates that Gray’s material three petals, that are usually emergent and response to cold water conditions, whilst was collected in south-eastern USA. The abundant perisperm in the seeds (Figures Ørgaard (1991) suggests shoot colour is plant is called cabomba or fanwort in Aus- 1 and 2) (Osborn et al. 1991). strongly influenced by light conditions; tralia and Carolina watershield, Washing- There has been some confusion over shoots becoming reddish brown in bright ton grass and fish grass in the USA the species constituting the genus light. Colour is the only morphological (Chittenden 1951, Mabberley 1987, Par- Cabomba. Eleven species have been de- trait that distinguishes C. caroliniana var. sons and Cuthbertson 1992). scribed: C. aquatica Aublet, C. australis pulcherrima (Wain et al. 1983). Because of Spegazzini, C. caroliniana A. Gray, C. the ecophenotypic plasticity, differentia- Description furcata Schultes & Schultes f., C. haynesii tion of Cabomba species is best done on the It is now generally agreed that Cabomba Wiersma, C. palaeformis Fassett, C. basis of seed characteristics such as size, and the related genus Brasenia constitute piauhyensis Gardner, C. pubescens Ule, C. shape and surface structure (Ørgaard a separate family, the Cabombaceae pulcherrima (Harper) Fassett, C. schwartzii 1991). On the basis of flower colour, three (Ørgaard 1991). The Cabombaceae is char- Rataj and C. warmingii Caspary. Five spe- varieties of cabomba are now distin- acterized by submerged rhizomatous cies are currently recognised: C. aquatica guished (Ørgaard 1991): C. caroliniana var. stems, floating, long-stalked, peltate Aublet, C. caroliniana A. Gray, C. furcata caroliniana, C. caroliniana var. pulcherrima leaves or submersed short-stalked, Schultes & Schultes f., C. haynesii Wiersma and C. caroliniana var. flavida with, respectively, white, purple and yellow flowers. The cultivars rosifolia, multipartita, paucipartita and ‘Silbergrune’ (= ‘Trifolia’) are traded by aquarists but all appear to be clones of C. caroliniana (Chittenden 1951, Rataj and Horeman 1977, Wain et al. 1988, Martin and Wain 1991, Ørgaard 1991). Unless otherwise specified, in this paper the term ‘cabomba’ refers to C. caroliniana as defined by Ørgaard (1991) and its horticultural varieties. Because of the confusion surrounding the identity of cabomba in Australia a detailed description is given here, based on Raciborski (1894, in Sculthorpe 1967), Fassett (1953), Sanders (1979), Schneider and Jeter (1982), Moseley et al. (1984), Ito (1986) and Ørgaard (1991). In the seedling the first few pairs of leaves are lanceolate and devoid of lami- nae. These are succeeded by transitional leaves until the submerged adult leaf form appears. Plants are strictly aquatic and com- Figure 1. Cabomba caroliniana showing the mass of submerged leaves, pletely submerged except for flowers and small floating leaves and solitary flowers (image provided by the occasional floating leaves. Stems may be Information Office of the University of Florida, IFAS, Center for Aquatic up to 10 m long. They are slightly com- Plants, Gainesville). pressed, 2–4 mm in diameter and increase Plant Protection Quarterly Vol.12(4) 1997 155 in width acropetally in the internodal re- alternate, with the blades peltately at- trichomes that secrete a gelatinous mucus gion. Scattered short, white or reddish- tached to the petioles (Figure 2) and they which covers the entire plant. The vena- brown hairs are present. The ‘rhizomes’ have a firm texture. Floating leaves are tion of the submerged leaves corresponds are erect, stout stems which have become peltate, entire, green to olive green, nar- to the leaf division. Erect and flowering prostrate and partially buried; they are rowly elliptic or ovate (trullate or shoots have proximately decussate not true rhizomes. They have opposite sagittate), 5–20 mm in length and 1–3 mm phyllotaxy which changes near the sur- buds and sometimes small leaves. Some wide. They are borne on the flowering face of the water to 1/3 with a flower at rhizomes are runners (horizontal) and branches. Submerged leaves are petiolate, nearly every node. possess upturned, erect heads. New rhi- opposite or less commonly in whorls of Flowers are solitary and raised above zomes and floating shoots arise as axillary three, and divided. The divisions are lin- the water surface (Figure 2) attached to a branches to these shoots. Rhizomes are ear but the terminal divisions are slightly long axillary stalk and are 6–15 mm in fragile and break and decay quickly. Ad- spathulate. The semicircular leaves may diameter and 6–12 mm long; milk-white, ventitious roots occur on the rhizomes at be divided dichotomously and trichoto- pale yellow or purplish. The flower is her- 45 degrees dextral and sinistral to the mously several times (Figure 2) to give a maphroditic and generally trimerous but leaves and branches. great number of terminal divisions: from di- and tetramerous flowers are found. Se- Shoots are grass green to olive green, 3–20 on the lower stem to 150–200 for pals three, elliptic to obovate, 5–12 × 2–7 sometimes reddish brown, often coated larger apical leaves. The leaflet margin is mm, pale yellow to milk white, if whitish with mucilage and more or less pubescent. serrulate to denticulate, the teeth being then often purplish tinged on margins and When present, the floating leaves are barely visible. The teeth are really 3-celled veins, base often greenish yellow on the

b c

a f g h

Figure 2. Cabomba caroliniana (Redrawn after Watson and Dallwitz 1995 onwards) (a) submerged leaves, floating leaves and flower, (b) flower, (c) pistil, (d) base of seed showing embryo enclosed by albumen, (e) fruit, (f) vertical section of carpel showing ovule attachment, (g) seed showing sculpturing and (h) L.S. seed. 156 Plant Protection Quarterly Vol.12(4) 1997 abaxial face. The three petals alternate with the sepals and are slightly fused to- ● gether at the base and are obovate to elliptic, 4–12 × 2–5 mm, pale yellow to white, purple tinged or bright purple, with apex ● obtuse or emarginate. The petal base is ex- ●● ● tended into two equal semicircular lobes ● curved more or less inwards towards the middle of the petal and partially covering the claw (Figure 2); the lobes with two more or less conspicuous yellow, elliptic, separate patches which function as nectaries; claw a deep yellow at the base, ● becoming paler apically. Stamens (3–)6 ● ● usually in one whorl. If six are present they are inserted on radii between petals ● and sepals. If three, they are antepetalous. ●● ● Carpels (2–)3(–4), antepetalous when ● four, divergent at maturity and with 1–5 ● anatropous, pendulous ovules (Figure 2) and stigma clavate. Two ovules are attached at the dorsal side and one at the ● ventral side. Seeds (Figure 2) ovate to el- lipsoid-oblong, 1.5–3.0 × 1.0–1.5 mm long, verrucate, the cristate-costate outer testal layer composed of irregularly digitate cells uniformly and densely perforated Figure 3. The known distribution of cabomba in Australia. with simple pits with four longitudinal rows of tubercles formed by the radial Oyster Bay, New South Wales. In Queens- native to the south-eastern USA and elongation of digitate cells. Seeds covered land, cabomba was first noticed as a pest southern Brazil, Paraguay, Uruguay, and with elongate processes and coated in a in 1989 when, as a result of an aquarium north-eastern Argentina (Ørgaard 1991). gelatinous slime. Pollen grains 60–90 µm escape, it was infesting the swamp that fed This distribution and that of other mem- in polar axis, prolate (boat shaped to ob- Leslie Creek (Atherton Tablelands) al- bers of the genus suggest it could be natu- long), elliptic, monosulcate, supratectal though it had been present since 1986. It ralized in the USA and originally be a na- sculpturing striate; monads at maturity. overgrew the fish breeding ponds at tive of South America. Fruit (Figure 2) green when fresh, carried ‘Quinkin Ponds’ and by the end of the Because of its extensive use in the below the water surface, apocarpous with year it had infested the length of the creek aquarium trade, cabomba has been intro- 1–4 dark brown carpels. and had spread into at least one arm of duced to Malaysia, India, Japan and New Chromosome number: 2n=39, c.78, Lake Tinaroo, into which the creek flows. Guinea (Ørgaard 1991). In Japan cabomba c.104; the basic number proposed by By 1991 further infestations were reported is considered a noxious weed (Oki 1992). Ørgaard (1991) is x=13 (not x=12 as previ- from Avondale Creek, north of Cairns and Cabomba is sold as an aquarium plant, but ously proposed; counts of 2n=24 are prob- a drainage channel at Goondi, near is not yet naturalized in the ASEAN re- ably miscounts) so specimens of cabomba Innisfail. In southern Queensland, con- gion (Malaysia, Singapore, Philippines, are apparently triploid, hexaploid or cern about its weed potential developed Indonesia, Thailand, Brunei) where it has octoploid. The ploidy of Australian when it was first observed in Six Mile the potential to cause serious problems in cabomba has not been reported. Creek, the original impoundment for Lake aquatic ecosystems. It is considered im- Seeds have not been reported from MacDonald (Noosa shire), in April 1992 perative that strict quarantine regulations cabomba in Australia, so current identifi- (Anderson and Garraty 1994) although are enforced against cabomba in the cation must be largely based on vegetative non-weedy outbreaks were observed in ASEAN region (Revilla et al. 1991). characters. the Caboolture River in 1991. The known distribution of naturalized As a result of the infestation in Leslie cabomba in Australia is shown in Figure History Creek, its spread into Lake Tinaroo and 3. Infestations are currently restricted to Cabomba is of relatively recent introduc- the possible infestation of associated irri- the Northern Territory, north-east tion into Australia; the earliest record is gation systems, cabomba was declared as Queensland (Atherton Tableland and from 1967 (Garraty et al. 1996) but it was P2 (to be eradicated) in Atherton and coastal districts) and the south-eastern not considered naturalized in Australia by Eacham shires in May 1990 and the decla- Queensland, central and northern New Sainty and Jacobs (1981) and was first ration extended in July 1992 to the South Wales coasts, with a single inland added to the New South Wales flora in Johnstone and Mulgrave shires. record from the Griffith area. In Victoria, 1986 (Jacobs and Lapinuro 1986). A long- To date C. caroliniana var. caroliniana is cabomba is largely restricted to South standing assumption has been that natu- the only taxon to have naturalized in Aus- Gippsland but it is considered a potential ralized populations are due to aquarists tralia, although pink cabomba (C. furcata, threat to permanent, freshwater, aquatic dumping unwanted plants into local wa- previously C. piauhyensis) and green vegetation throughout the state. Currently terways, but more commonly, areas have cabomba (C. caroliniana var. caroliniana, it is rare or localized although some been deliberately planted to allow wild previously C. australis) are regularly populations are quite large (Carr et al. cultivation for the aquarium trade. traded by aquarists. 1992). Cabomba was introduced into Aus- The potential distribution of cabomba tralia from the USA as an aquarium plant. Distribution in Australia is difficult to establish as there The earliest Australian herbarium record Cabomba has a curious and disjunct dis- is no good model available for predicting is in 1967 from cultivated material at tribution in that it is considered to be the distribution of a fully aquatic weed. Plant Protection Quarterly Vol.12(4) 1997 157 However, two approaches have been used. Rules for cabomba were based on zone (Figure 3) which may mitigate taken. Firstly, CLIMEX (Skarratt et al. average air temperatures and their ranges, against this problem. 1995) has been used to model the weed’s average rainfall and rainfall variation, soil The CLIMEX model was developed by potential distribution based on tempera- type and soil nutrients. Clearly, with a matching the USA distribution of ture tolerance, since temperature is a ma- weed such as cabomba which is in its ini- cabomba. The final model predicted the jor determinant of distribution in aquatic tial invasive stages, all environments in distribution of cabomba in the south-east- plants (Sculthorpe 1967). Secondly, GARP which it can occur may not be represented ern USA but also predicted the presence (Stockwell 1996a,b), a rule-based model in its current distribution and this limits of the weed along the west coast. This that deduces from a species’ current Aus- the usefulness of this second approach. probably indicates that the distribution in tralian distribution environmental rules However, in Australia, cabomba is cur- the USA is not determined solely by tem- which determine its distribution has been rently distributed across a wide climatic perature. Interestingly, subsequent to the development of this model, the first ●● ● records of cabomba in this region (Wash- ●●● ● ington and Oregon) were found (Anon. ● ● ● ● ● 1995). The model predicted the recorded ● ● occurrence of cabomba in South America ● ● ●● ● (Ørgaard 1991) but also a more extensive ● ●¡ distribution, suggesting again that tem- ● perature is not the sole factor determining ●● the distribution of this species. The model ● ● ● ¡ ● ●● ● suggests that much of coastal Australia, ● ● except for the north-west, is excellent ●● ● ● ● ● ● ● ● habitat for cabomba, and the optimal area ● ●● ● ● ● ●¡ ● ● for growth of the weed is coastal Queens- ● ● ● ● ● ●¡ ● ● ● ● ● ●¡ land (Figure 4). ● ●● ● ● ● ● ● ● ● ● ●¡ GARP predicted a much more re- ● ● ● ● ● ● ●● ● ● ● ● ●● ● ●● ● ●●●● stricted distribution of cabomba in Aus- ● ● ● ● ● ● ●●● tralia (Figure 5) than CLIMEX, and indi- ●●●● ● ●● ● ●●● ●●●●●● ●●●●● ● ● ● ● EI cated the distribution would be centred on ● ● ●● ● ●● ● southern coastal Queensland and New ● ● ●● ● ● 10 ●●●● ● ● South Wales. ●● ●●● ● CLIMEX probably overestimates the ● ● 25 ● ● potential distribution of cabomba whilst ●● ● 50 GARP probably underestimates its poten- ●● ● 75 tial distribution. Both models make pre- dictions of the distribution irrespective of ¡ 100 whether there is a suitable water body for Figure 4. The distribution of Cabomba caroliniana in Australia predicted the weed to inhabit. Nonetheless, the results from both models suggest that the using CLIMEX (EI=Ecoclimatic Index; EI=10, potential for a permanent southern coastal strip of Queensland is population extremely low; EI=100, this potential extremely high). suitable, if not optimal, habitat for cabomba.

Habitat Cabomba is found in ponds, ditches, small shallow lakes and slow flowing streams in coastal vegetation of swamp forest and bog, and inland in areas of savanna (Ørgaard 1991). It can grow in water up to 10 m deep but most commonly the plants grow rooted in shallow water (up to 3 m) (Tarver et al. 1978, Sanders 1979, Schnei- der and Jeter 1982, Hanlon 1990). Ørgaard (1991) suggests that cabomba species are adapted to habitats subjected to significant annual water level fluctua- tions through having the ability to grow rapidly to keep pace with increases in water depth during the wet season. If so they would be pre-adapted to conditions found in many Australian aquatic systems. Like most fully aquatic plants, cabomba is sensitive to the drying out of its habitat. In experiments conducted in aquaria, only 6.7% of cabomba seedlings survived a 30 day drawdown of the water level in which the hydrosoil remained unsaturated, and Figure 5. The distribution of Cabomba caroliniana in Australia predicted growth of survivors started within 14 using GARP. days of refilling aquaria. If the hydrosoil 158 Plant Protection Quarterly Vol.12(4) 1997 was saturated, 53% of seedlings survived turbid and turbid water caused by inflows Perennation and regrowth was evident after seven usually helps to control aquatic weed Cabomba is an aquatic perennial (Tarver days (Sanders 1979). These observations problems, this characteristic of the weed is and Sanders 1977), growing from short support Ørgaard’s suggestions. of concern. rhizomes with fibrous roots (Hellquist and Crow 1984, in Madsen 1996). Climatic requirements Light Cabomba likes a warm-temperate, humid In contrast to the above findings, in culti- Physiology climate, with rain throughout the year and vation, cabomba is demanding of light Cabomba is not cyanogenic; it contains an annual temperature of 15–18°C and water quality and sensitive to compe- alkaloids, but saponins/sapogenins are (Ørgaard 1991). Although it can with- tition and water motion (Ørgaard 1991), absent (Watson and Dallwitz 1995). stand temperatures of less than 0°C, its although Anderson and Garraty (1994) preferred temperature range is 13–27°C state that as an indoor water plant it toler- Phenology (Leslie 1985). ates poor lighting. Little is known of the life cycle of cabomba in Queensland. In north Queensland it pH Substratum grows and flowers continuously through- Cabomba is reported from acid and alka- In Queensland cabomba appears to prefer out the year, and in south-east Queens- line waters (Ørgaard 1991), but the opti- silty substrata into which it does not root land buoyant stems up to 6 m long are mum pH for growth is 4–6 and growth in- deeply. Where it occurs on hard or stony produced in summer with a growth rate hibition occurs at pH 7–8. Above pH 8 the substrata the plant’s vigour is reduced of up to 5 cm per day (P. Bell personal stem becomes defoliated and growth is in- (Garraty et al. 1996). communication 1995). In July and August hibited (Riemer 1965, Gregory and Sand- these stems lose buoyancy, lie across the ers 1974, Tarver and Sanders 1977). Con- Plant associations surface of the hydrosoil and fragment. sequently cabomba grows best in acidic In south-east Queensland cabomba co- When growth recommences, fragments waters (such as those around the Florida exists with (but tends to dominate) sub- reroot and initiate new plants (G. Diatloff panhandle) (Hanlon 1990). This is perhaps merged native species including Hydrilla personal communication). These observa- accounted for by nutrients being more verticillata (L.f.) Royle, Nymphoides indica tions are similar to those of Riemer and available in acidic waters (Gregory and (L.) Kuntze, Eleocharis sphacelata R. Br., Ilnicki (1968). In mild winters, this die Sanders 1974, Sanders 1979). Hygrophila salicifolia (Vahl) Nees, back may not occur (T. Anderson personal Polygonum spp., Najas tenuifolia R. Br., communication). Nutrients Potamogeton javanicus Hassk., Utricularia The following description refers to the In Japan cabomba grows well in nutrient spp., Vallisneria spiralis L. and the exotic life cycle of cabomba in the USA. Towards rich water. The ranges (mg L-1) of habitat Nymphaea capensis Thunb. (Anderson and the end of the growing season, stems be- variables in which it occurs in Japan are: Garraty 1994). Cabomba grows in associa- come denuded and brittle and hard. Ter- COD 3.2–8.23, inorganic-N 0.68–1.76, and tion with Lepironia sp. and salvinia minal stems especially tend to break free organic-N 0.06–0.25 (Oki 1992). Cabomba (Salvinia molesta D.S. Mitchell) in the and these remain green and leafy until also grows optimally at very low calcium Glenbrook Lagoon in New South Wales. spring. Some terminal buds remain at- ion concentrations (4 ppm); higher levels In the eastern USA cabomba coexists tached to the substratum, even under ice. of calcium inhibit growth (Riemer 1965). with other native aquatic herbs including Growth starts around April (Riemer and Nymphaea odorata Aiton, Nuphar lutea (L.) Ilnicki 1968). Even defoliated stem frag- Turbidity Sibth. and Smith, Myriophyllum spp. in- ments buried in mud under ice may re- The response of cabomba to water turbid- cluding M. aquaticum (Vell.) Verdc. and grow (Ørgaard 1991). ity has been investigated in aquaria Brasenia schreberi J.F. Gmelin (Guerra 1974, Most of the material supplied to the (Gregory and Sanders 1974, Sanders Schneider and Jeter 1982). The latter au- aquarium trade originally came from the 1979). Growth was measured at low thors report that it is extremely persistent, San Marcos River in Texas. Here the plant (30–45 Jackson turbidity units) medium forms dense stands and under good grow- is perennial and flowers throughout the (70–110 JTU) and high (300–2350 JTU) ing conditions can crowd out other previ- year, and although the number of flowers turbidities. Growth at medium turbidities ously established native species. drops drastically during November– was greatest, followed by growth at high February, isolated flowers can still be turbidities. Moderate turbidity (70–110 Growth and development found. Floating leaves are produced dur- JTU) enhanced stem length growth com- ing the flowering period (Schneider and pared to non-turbid (0–10 JTU) conditions Morphology Jeter 1982) and fruits appear shortly after and it is postulated that this is due to an The general morphology of the plant is the first flowers have emerged (Riemer auxin effect, producing longer cells at shown in Figures 1 and 2 and has been and Ilnicki 1968). Fruit take about a month moderate turbidity levels. Moderate to described by Ørgaard (1991). One to 10 to mature (Schneider and Jeter 1982). high turbidities (300–350 JTU) enhanced adventitious roots can form spontane- In the southern USA peak seed produc- adventitious root development. De- ously at the nodes of erect, vegetative and tion is during May–October, but occurs creased underwater light intensity gener- free floating shoots, without the need for from April to December (Sanders 1979). In ally leads to growth limitation in sub- contact with a substratum. In Queensland, Louisiana cabomba flowers from May to merged aquatic plants, so this result is 3–40 strong, flexible stems, 2–6 mm thick, December, with peaks in May and Sep- somewhat counter-intuitive. In these ex- arise from a single root mass. Some hori- tember (Sanders and Mangrum 1973). In periments turbidity was maintained by zontal stems may become buried and frag- New Jersey, it begins flowering in late stirring the hydrosoil. This could have led ment to give rise to new erect shoots on June/early July and maximum flowering to an increased release of nutrients to the separate plants (Garraty et al. 1996). Ad- occurs in late July and continues through water, hence increased availability to the ventitious roots are thin, white and August. A few flowers continue to emerge plant, as the shoots and stem are the main unbranched. Older, embedded roots are until the first frosts (Riemer and Ilnicki sites of nutrient uptake (Sanders 1979). numerous, long, very slender, branched 1968). Fruit set is often abundant through- Nevertheless, cabomba appears to be able and purple when fresh, drying to dark out most of the year, but most seed is pro- to grow well in turbid conditions, and brown to black. duced between May and October (Hanlon since Australian freshwaters are generally 1990). Plant Protection Quarterly Vol.12(4) 1997 159 Mycorrhiza Schneider and Jeter (1982) claim that the seeds are released 14–30 days after There are no records of mycorrhizal asso- autogamy and apomixis do not occur and anthesis. The seed is green when fresh ciations with cabomba. protogyny is absolute (stigmas only recep- with a scattered dark pigmentation and tive on the first day and never on the sec- slowly turns brown with age. It is globose Reproduction ond day) but Ørgaard (1991) observed to ovoid-oblong with slightly flattened fruit setting without either hand pollina- ends (Ørgaard 1991). Seed anatomy is Floral biology tion or detected insect pollination, so that similar to that of the nymphaceous genera: This account of the floral biology of some degree of autogamy seems possible. there is abundant perisperm, little cabomba is based on Tarver and Sanders Similarly, Sanders’ (1979) observations endosperm, a haustorial tube and a small (1977), Sanders (1979), Schneider and Jeter suggest that cabomba is only facultatively dicotyledonous embryo (Schneider and (1981, 1982), Ørgaard (1991) and Osborn entomophilous. Wind or rain would be Jeter 1981). et al. (1991). sufficient to displace pollen onto the stig- Schneider and Jeter (1982) indicate that Pollen fertility is low in cabomba mas. Certainly though, self-pollination submerged as well as emergent flowers (45–95%) compared to the rest of the ge- appears to be a rare event (Hanlon 1990). are produced in the San Marcos River, but nus and is related to the high level of Tarver and Sanders (1977) found water, pollen release does not occur in these polyploidy. Flowers contain relatively wind and hand pollination failed to pro- flowers and seeds are not produced. few, large pollen grains and have low pol- duce seed but that insect pollination was Seeds have not been recorded from len-ovule ratios (560 ± 123 [95% confi- successful. Forty per cent of flowers which Australian plants although two her- dence limits] grains per flower, pollen- had been visited by insects produced ma- barium specimens from south-east ovule ratio 62 ± 14). These features are ture seed after about one month. Twenty Queensland possess fruit. Since potential characteristic of entomophily (insect polli- per cent of flowers visited on the first pollinators are plentiful, cabomba in nation). Flowers undergo dianthesis over day produced seed and 60% of flowers Queensland may be sterile. two consecutive days, during which flow- visited on the second day produced seed, Ørgaard (1991) suggested that as for ers are structurally and functionally pistil- so pollination is due to insects and cross- most water plants, seed dispersal is ef- late on the first day and staminate on the pollination is the rule. Principal pollin- fected by water birds, and this could be second. Flowers open in the morning ators were Enallagma and Anax (Odonata), important in ensuring dispersal between around 10 a.m. and close in the afternoon Halictus and Apis (Tarver and Sanders lakes or river systems. Sanders (1979) at around 4 p.m. On closing, flowers sub- 1977) but Odonata probably are acciden- states that the floating pistil helps disperse merge. When open they are raised 1–4 cm tal pollinators as they are not nectar or the seed. However, cabomba’s range ex- above the water surface due to elongation pollen feeders. The major pollinators ob- tension in the USA is generally considered of the peduncle. This occurs some two served by Schneider and Jeter (1982) were to be due to the discarding of unwanted hours before anthesis. Consequently, dur- small ephydrid flies. plants by aquarists (Hanlon 1990, Madsen ing the morning, second day flowers Cabomba flowers are visited by numer- 1996). stand slightly higher than first day flow- ous insects, particularly small flies. On the ers. first day whilst a fly is taking nectar, a Physiology of seeds and germination Flowers are protogynous on the first stigma is in close proximity to its back or In the colder, north-western part of day, stamens are short filamented and head. On the second day the anther is in cabomba’s USA distribution, fruit setting indehiscent, and the longer pollen recep- this position. Flies were seen to carry pol- is often sparse (Ørgaard 1991) and in New tive stigmas arch out over the nectaries. len. The changes in morphology during Jersey, sexual reproduction is negligible The filaments elongate on the second day flower development ensure that pollen is or non-existent as Riemer and Ilnicki so that the anthers are level with the stig- transferred from the two-day-old flowers (1968) found no seedlings, no seeds ger- mas, but they point out towards the to the stigma of one-day-old flowers. Af- minated and no seeds were found with nectaries. The anthers undergo extrorse ter anthesis the flower is pulled beneath embryos. dehiscence on the second day some two to the water surface, either by an acute bend In Louisiana seed is produced but vi- three hours after the flower is fully open, in the peduncle, by coiling of the peduncle ability is low and only about 25% of seeds but by then the stigmatic papillae are flac- or loss of turgor pressure in the peduncle. germinate naturally (Sanders 1979). cid (suggesting loss of stigmatic receptiv- Only fertilized flowers (swollen carpels – About 5% of seeds germinate immediately ity) and the carpels re-oriented inwards. evident one week after pollination) were and do not require a period of after-ripen- Initially the pollen is a sticky mass, but it pulled underwater by the peduncle. Flow- ing. In experiments to assess the impor- dries and become powdery. Flowers wa- ers, not dehisced on the second day or wet tance of different environmental condi- terlogged by rain do not release pollen. If by rain, are pulled underwater but become tions on germination, only 1.8% germina- skies are overcast, anther dehiscence may water logged and do not release pollen. tion occurred. Seeds generally germinate be delayed by a few hours and stigmatal Abscission of the fertilized flower is pre- 5–10 weeks after fertilization (Tarver and sensitivity extended. vented by auxins released by the ovary Sanders 1977) but seed can remain viable The perianth persists until the seeds are and transported down the peduncle. The for more than two years. Factors believed released. Seeds are contained in long pis- coiling of the peduncle may also be due to to be important in affecting germination tils and vary from one to three per pistil auxins. The coiling is thought to protect are red light, temperature and high carbon depending on time of year (1 in May, 2–3 carpels from being broken from the pe- dioxide levels (Sanders 1979). The embryo in September). Fruit are mature by 2–4 duncle by fish and severe wave action. remains viable for up to eight hours if al- weeks. Pistils containing mature seed lowed to desiccate. Seed set in cabomba is separate from the pedicel and fall to the Seed production and dispersal reduced compared to that of congeners. bottom. The fruit wall decomposes and After fertilization, the perianth encloses This could be explained by reduced pollen the seeds are released and lie on the the fruits and within a few days the fertility and/or reduced stigmatic recep- hydrosoil surface. Eventually, seeds sink anthocarp becomes submerged due to re- tivity; both associated with high ploidy. into the substratum and are here protected curvation of the pedicel (Ørgaard 1991). Climate and the environment may also af- from desiccation. It is speculated that Fruit (pistils) break away from the plant fect seed set in this species (Schneider and seeds have the capability of remaining vi- and fall to the bottom, decompose and Jeter 1982). able after long periods of desiccation or leave the seed at the hydrosoil surface dormancy (Madsen 1996). (Sanders 1979). The perianth persists until 160 Plant Protection Quarterly Vol.12(4) 1997 Vegetative reproduction competitive and can exclude well estab- can interfere with recreational, agricul- Cabomba grows and disperses mainly lished native species (Riemer and Ilnicki tural and aesthetic functions of lakes and from fragmentation (Sanders 1979, 1968), it can itself be out-competed by such reservoirs (Riemer and Ilnicki 1968). In the Hanlon 1990). Any detached shoot with at weeds as Egeria (=Anacharis) densa Planch. USA commercial fishing camps have been least one pair of expanded leaves is capa- (Sanders and Mangrum 1973). Extracts of forced to close or have had their income ble of growing into a mature plant. Larger cabomba have allelopathic effects at me- severely reduced (Sanders 1979). sections than this can root at the nodes. A dium and high concentrations and since Similar losses are unlikely to occur in motorboat passing through a cabomba allelopathy plays a role in determining Australia since natural lakes are relatively bed can produce hundreds of dissemin- the distribution of higher plants (Rice few and impoundments are not as inten- ules and in many situations this is prob- 1979), this could be a mechanism whereby sively utilized. Nevertheless, the potential ably the major dispersive and infestive cabomba can oust native species exists for economic losses from damage to mechanism (Sanders 1979). (Elakovich and Wooten 1989). amenity values. Even threats to human As Riemer and Ilnicki (1968) note, veg- health and safety may ensue as water ski- etative reproduction is very important in Importance ers or swimmers could easily become en- many aquatic plants. For example, tangled by the weed’s long thick stems Ceratophyllum demersum L. often does not Detrimental (Figure 6) and drown. reproduce sexually because conditions are Cabomba is still in its early invasive stages In Queensland, cabomba infestations too cold, but reproduces entirely by axil- in Australia and the actual extent of the may deleteriously affect water quality lary buds on plant fragments from the pre- infestation is not known. Currently it ap- (Anderson et al. 1996, Garraty et al. 1996) vious year. Nevertheless, the species is pears to be having little impact but until a through increasing water colour, with a abundant and often dominant in fresh- survey has established its detailed distri- subsequent estimated increase in the treat- waters where it does not sexually repro- bution, its true importance cannot be ment costs for potable water of $A50 per duce. known. Meanwhile, its potential impact ML. There is also a suggestion (T. can be judged by reference to its impact Anderson personal communication) that Hybrids overseas. in southern Queensland the sudden re- Reproduction in cabomba is largely Cabomba has the potential to cause lease of manganese caused by the dieback asexual and hybrids have not been re- blockages in the Panama Canal (Hearne and decomposition of large amounts of corded. However, since C. haynesii is pos- 1966). However, most information on cabomba in the winter months could affect sibly a hybrid between C. palaeformis and cabomba is from the USA. It is a problem the manganese cycle and cause a reduc- C. furcata (Ørgaard 1991), hybrids involv- throughout the Gulf states, particularly in tion in water quality. Further research is ing cabomba could conceivably occur. Louisiana (Tarver and Sanders 1977). In required into these situations. If present in Florida, although the plant is extending its water storages, heavy infestations, be- Population dynamics range (Sanders 1979) and increasing in cause of the large volume of plant mate- Cabomba is capable of rapid spread once abundance, it is not yet regarded as a nui- rial, could cause water loss from overflow it has been introduced into a suitable wa- sance plant (Hanlon 1990, Martin and or seepage. Owing to these problems and ter body. It was first reported from Lake Wain 1991). It is, however, one of the 19 the ability of cabomba to grow rapidly, MacDonald (south-eastern Queensland) plant species that cannot be transported, cabomba has the potential to become a in April 1992, and by 1995 had invaded imported, cultivated, collected or sold in major weed in water storages. almost the whole of the lake’s extensive Florida (Clugston 1990). Irrigation canals could provide an ideal littoral zone, where it had replaced much Heavy infestations can raise water lev- habitat in which cabomba could grow and of the submerged native vegetation els causing overflows and seepage losses. where it could impede water flow, cause (Anderson et al. 1996). Although under Oxygen depletion can occur when mas- overflows and blockages. Although it is suitable environmental conditions sive dieback and consequent decomposi- difficult to assess the economic loss this cabomba is extremely persistent and tion occurs (Gracia 1966). Dense stands might cause, the nuisance value could be high. Anderson and Garraty (1994) have assessed the impact of cabomba on native aquatic plants and water quality in Lake MacDonald, Queensland. In summer the mean standing crop of cabomba was 1.02 kg m-2, a seven fold increase from early spring levels. Other species were found only at very low standing crops in infested areas, compared to their abundance in uninfested areas. These results suggest that a primary concern with cabomba in Australia should lie with its potential as an environmental weed. Its ability to re- place native aquatic plants, with the likely displacement of native fish and invertebrate populations, together with the ability to infest large areas of water, suggest that native aquatic life would be consider- ably endangered if cabomba was allowed to establish throughout the country. Control costs are currently minimal as very little control has been attempted and Figure 6. Cabomba showing the long thick stems which could easily most of this has been associated with ex- entangle divers or swimmers and impede boating and water flow. perimental trials. In Queensland so far, Plant Protection Quarterly Vol.12(4) 1997 161 $A250–300 000 has been spent in trying to also note that all species of cabomba are Sonar® (fluridone) is an effective herbi- control cabomba in the Ewen Maddock sold as aquarium plants. They are there- cide for cabomba (Tarver 1985). It gives Dam (R. Rainbird personal communica- fore likely to escape into suitable freshwa- long term control, is easy to apply and is tion) and estimated costs for mechanical ter habitats in countries to which they are selective, as many plants are not suscepti- control in Lake MacDonald are $A125 000 introduced. Due to their similarity with ble (Tarver 1985, 1987). It is not volatile, is for the harvester and $A20 000 per annum C. caroliniana and the ease of vegetative unaffected by pH (4–14), has an average for harvesting costs (K. Garraty personal propagation, it is possible that those spe- half-life of 20 days and is decomposed by communication). The cost of treating cies not already commercially available in ultra-violet light into non-herbicidal, non- a volume of 20 000 m3 of water Australia will become so, and that species toxic products. It is not deactivated by ad- (100 × 100 × 2 m) with the 2,4-D n-butyl other than cabomba will become natural- sorption onto suspended organics or clay ester/diatomaceous earth mixture (see be- ized. particles, hence it is effective in turbid wa- low) would be approximately $A3000. ters. It acts by interrupting carotenoid syn- Clearly, on cost grounds alone, it is un- Weed management thesis and since carotenoids seem to pro- likely large scale infestations would be tect chlorophyll from photo-degradation chemically treated. Herbicides by UV light, affected plants show symp- Whilst it has been reported that cabomba toms of bleaching (chlorosis). Fluridone is Beneficial is susceptible to a variety of commonly a systemic herbicide and plants will ab- Where naturalized, cabomba provides the used herbicides (endothall, 2,4-D, 2,3,5-T, sorb it via the leaves and shoots and the usual benefits that aquatic plants gener- silvex, diquat, dichlorprop), their effects roots. Control normally takes 30–90 days ally have in aquatic systems: it oxygenates are erratic (Hiltibran 1974, 1977), retreat- and this slow action helps prevent oxygen the water, protects against bank and bed ment is often necessary and managers depletion of the water due to massive de- erosion and removes nutrients from the consider cabomba difficult to control composition of dead vegetation. Control is water. It can also provide cover for young (Leslie 1985, Madsen 1996). best achieved during periods of active fish and a habitat for invertebrates, as well Results of trials with 2,4-D are incon- growth. To control cabomba in Florida, as being a source of food for wild life, in- sistent (Hiltibran 1974), but a granulated but to avoid damage to other susceptible cluding water fowl (Oki 1992). However, formulation of 2,4-D as the butoxyethanol plants such as lilies, a very early spring whilst it does provide fish habitat in the ester has been useful as a treatment application (January–March) is used. Tox- USA, it has no wildlife value (Martin and for water weeds, including cabomba, in icity to fish occurs at about 7.6–22 ppm Wain 1991). In regions where it is inva- potable water supplies in Texas (Guerra and to invertebrates at circa 1.4–4.4 ppm sive, it is not clear whether native fish and 1974). (the normal application rate concentration invertebrates utilize it readily as a habitat. Symmetrical triazines (e.g. simazine) is around 0.1 ppm). The US Environmen- Research is needed to clarify the situation are well known as effective aquatic herbi- tal Protection Authority has concluded in Australia. cides. The growth of cabomba has been that Sonar• does not pose a risk as a Like many aquatic weeds, cabomba ef- reduced by treatment with the triazine chronic or acute toxicant in aquatic sys- fectively removes plant nutrients (phos- terbutryn. However terbutryn can stimu- tems. Only the slow release formulation phorus and nitrogen) from water. late growth at low dosages (Riemer and (release occurring over 7–14 days) is ap- Anderson et al. (1996) showed cabomba Trout 1980). proved for rivers by the US EPA and con- reduced dissolved nitrogen by 25% and The potassium salt formulation of trol is poor if it is applied during rapid dissolved phosphorus by 44% in Lake endothol-silvex controls cabomba, par- flow conditions. In still waters a spreader/ MacDonald. Harvesting and removal of ticularly with the addition of surfactant sinking agent is recommended (Tarver cabomba may therefore be a way of limit- (X-77®), and has a low toxicity to fish and 1987). When used in northern Queensland ing eutrophication in some waters. mammals (Lapham 1966). The granular on Lesley Creek in an experimental field Cabomba is a widely grown and com- formulation is recommended for the mar- application, cabomba control was ineffec- mercially important aquarium and out- gins of deep water areas. tive, perhaps due to the application being door aquatic plant in many countries. Its Endothall acid can be used to control into moving water. Sonar is not commer- finely dissected submerged leaves are at- cabomba if formulated as the potassium cially available in Australia. tractive (particularly in the purple form) salt as in Aquathol K®, or the more toxic In south-east Queensland before 1992, and it is used as an aquarium oxygenator alkylamine salt as in Hydrothol 191®. For no attempt had been made to control or (Chittenden 1951, Mabberley 1987). many years Hydrothol 191 has been used eradicate cabomba infestations because it Cabomba grows well in shallow, well lit to control cabomba in Florida at applica- had appeared non-invasive and there and eutrophic freshwater, flowering pret- tion rates of 0.5–1.5 ppm without causing were no suitable herbicides registered for tily at the surface and rapidly forming fish kills or obvious adverse environmen- use. With the advent of the weed in two large colonies. tal effects. Endothall acid breaks down water storages, it was realised effective rapidly and completely. It does not leave control methods were required. As a re- Legislation residues, accumulate in the hydrosoil or sult an effective and relatively cheap Cabomba is proclaimed as a prohibited or food chain, or move significantly from the chemical control for cabomba has been restricted plant only in Queensland, treatment site. Emergent plants need not devised (Diatloff and Anderson 1995). where all Cabomba species are declared be affected and it is not toxic to fish. Not 2,4-D n-butyl ester plus diatomaceous plants under the provisions of the Rural more than 10% of the water body should earth is mixed at 1 part to 20 parts of water Lands Protection Act (1985). The genus is be treated at any one time with rates ex- and injected 2 m below the water surface declared as category P3 for the whole ceeding 1.0 ppm. A weighted hose should through a series of weighted nozzles to state: where found, the numbers or distri- be used to apply the liquid herbicide as achieve a final concentration of 10 ppm bution of the plant should be reduced. It is close to the bottom as possible, or prefer- clay/2,4-D active ingredient. This method illegal to sell or keep the plant throughout ably, the granular formulation should be of application allows the mixture to the state. used (Moore 1991). spread sideways to provide a blanket Although cabomba is not declared, pro- Endothall is not registered in Australia cover of the area being treated and com- claimed or restricted in any other state, the except as a defoliant for crops and as a pletely cover the plant, which takes up the modelling results suggest that this would post-emergence herbicide for Poa annua L. herbicide through its leaves and stems. It be appropriate. Sainty and Jacobs (1994) (winter grass) in turf. is important that the approximate depth 162 Plant Protection Quarterly Vol.12(4) 1997 of the area being treated is known so that effective. If drawdown is used as a man- were noticed which were brought under the volume of water to be treated can be agement technique, water removal should control by hand pulling plants. estimated to allow the correct application be complete (Sanders 1979). rate. In trials in the Ewen Maddock Dam For several years drawdown has been Other treatments this method provided effective control of used as a management tool for aquatic Mechanical control methods can be very cabomba within a matter of days and weeds in Louisiana and is considered the effective for aquatic weeds and are the 2,4-D is now registered in Queensland for only economic method (Manning and most popular control method in Japan, but use against cabomba. Sanders 1975). Drawdowns of 1.5–2.5 m only temporary control is provided and Fluridone was also assessed as a herbi- have given 90% control of many weeds, this is expensive (Oki 1992). In the USA cide for cabomba in these trials but gave including cabomba, but unfortunately mechanical techniques have proven inef- almost no control. The reasons for this have enhanced the spread of water hya- fective (Madsen 1996). McComas (1994) clearly require more research, but cinth and alligator weed. Tarver and provides a comprehensive survey of me- fluridone is totally water soluble and in Sanders (1977) showed that consecutive chanical control methods, many of which the trials it appeared to dissipate com- autumn-winter drawdowns yielded a 99% could be used against cabomba. Cabomba pletely throughout the water body before reduction of cabomba in Black Lake, Loui- does not root deeply and can easily be any control could be effected. This may siana but cabomba re-established after the lifted out by the roots, although in deeper also explain why previous attempts at lake refilled, due to the germination of water this operation has to be carried out control using 2,4-D have given erratic re- seedlings. In Louisiana a complete winter by divers. A suction dredge has been de- sults. A water soluble form may have been drawdown is the best way to manage vised for use in the Ewen Maddock Dam used. In contrast 2,4-D ester is emulsifi- cabomba, although results are dependent (P. Bell, R. Rainbird personal communica- able, not water soluble and is adsorbed on weather conditions during drawdown tion) and a mechanical harvester has been onto the diatomaceous earth, so it does not (Manning and Sanders 1975). used for control of cabomba in Lake disperse very far. This enables quite pre- In the USA the use of drawdown has MacDonald (Garraty et al. 1996). The cise control in applying the herbicide and been controversial, owing to the economic mechanical harvester used in Lake localization of the area being treated. The losses ensuing from the temporary depri- MacDonald effectively halved the 2,4-D is quite selective in its action: other vation of fishing rights, boating facilities cabomba standing crop (from 48.7 to 25.9 t species present during the trials were not and hunting. Since the effective use of ha-1) but in three weeks cabomba had affected except for water lilies, which drawdown depends on the weather dur- grown back to pre-cut levels (51.9 t ha-1). quickly re-colonized from adjacent areas ing the drawdown period, these losses Cabomba grows well in nutrient rich (Diatloff and Anderson 1995). may not be balanced by effective weed waters and is an efficient utilizer of dis- With careful application and attend- control, so proper timing is essential. Dur- solved phosphate. In these situations har- ance to ensuring the required concentra- ing drawdown, cabomba fragments capa- vesting of cabomba may lead to an in- tion of active ingredient is met, it may also ble of growth may get carried or washed crease in water quality due to a reduction be feasible to use this method to control into the shallow remaining areas, and be- in dissolved phosphate and nitrate in the cabomba in slowly flowing streams and it come rooted. These areas then act as ref- water (Anderson et al. 1996, Garraty et al. should certainly be applicable to still wa- uges for the weed, resulting in rapid 1996). The removed material may be used ter canals. In the case of infestations in reinfestation (Sanders 1979). for composting, but if the amounts re- irrigation canals, problems with using In Australia, economic losses from moved are quite small, this may not be a 2,4-D could arise as many crops are sus- drawdown are not the same as in the USA, financial proposition, nor may it be neces- ceptible to the herbicide. However, if the as water bodies are not so intensively sary as cabomba placed on the bank de- canal can be bypassed, isolating the infes- used for recreational purposes. However, composes in 3–4 weeks. tation, and its use withheld for a period, water, particularly potable water, is a A problem with using mechanical con- control work can be carried out. Before scarce resource in Australia and draw- trols against cabomba is that the plant eas- bringing the canal back on line, bioassays down may not be acceptable for this rea- ily fragments and these fragments can can be performed using very sensitive test son. If it is an acceptable treatment, float away and recolonize the treated area plants to ensure no subsequent damage to drawdown is the best available option for or invade adjacent non-weedy areas. As a crops when the canal is brought back on- cabomba control, particularly in potable consequence, mechanical harvesting is not line. Since the half-life of 2,4-D in most water supplies. Generally the thick silt in suitable for small or new infestations, but aquatic systems is quite short, canals may which cabomba becomes rooted takes a may be the only acceptable method for only need to be off-line for relatively short long time to dry, so in the wet tropics large infestations in potable water sup- periods. where much of the current Queensland plies. Fragmentation is minimized by us- The use of 2,4-D in potable water sup- infestation is found, drawdown would be ing a suction dredge and, additionally, the plies may cause some concerns. However, best carried out in winter. If drawdown is whole plant is removed, including the when properly used, it is non-persistent in used, it may require supplementary treat- root ball. the environment at harmful levels and ments to guarantee weed control. Diuron For infestations in small creeks and ir- does not accumulate in food chains can be sprayed on the exposed root bases rigation canals, control through shading (Gangstad 1986). If used away from the to enhance and speed control and it is reg- may be viable, although cabomba does ap- take-off points, or if the reservoir can be istered in Queensland for use in this type pear to grow well at low light intensities. taken off-line for a while, with close moni- of situation. Alternatively 2,4-D n-butyl Adequate bankside vegetation can pro- toring, its use for the control of cabomba ester can be used. vide sufficient shade to stop submerged may be acceptable. In Queensland, two drawdowns (of 4 aquatic plants from growing (Dawson and 7 m respectively) have been used to 1989). If the weed beds are localized, a Drawdown control cabomba in the Ewen Maddock temporary covering of the affected area by Laboratory experiments have shown that Dam. The first gave incomplete control black shading fabric can effectively con- drawdown, with subsequent drying of the due to heavy rains partially refilling the trol the plants (Dawson 1989). This option hydrosoil, may be an effective manage- dam. The second drawdown was more may be particularly suitable for infested ment tool. In the field, extreme tempera- successful as the dam was dry for 4–5 irrigation canals, although weed frag- tures accompanying the drying out are months and most plants were killed. With ments must be contained to avoid infesta- likely to render drawdown even more partial refilling, two small reinfestations tion away from the treated area. Plant Protection Quarterly Vol.12(4) 1997 163 Natural enemies until the affected area is quite large. Sec- References Chinese grass carp (white amur, ondly, it grows very quickly and is highly Anderson, T. and Garraty, K. (1994). Ctenopharyngodon idella (Val.)) is an effec- invasive. Unless early control is initiated, Cabomba caroliniana downunder in Lake tive biological control agent for cabomba. the weed quickly establishes throughout MacDonald. 3rd Queensland Weeds It has been used in Arkansas, apparently the system and eradication is a hopeless Symposium, Toowoomba. Additional with no adverse effects on fish and water- task. paper. Weed Society of Queensland Inc. fowl populations. It is very effective, as the The first constraint could be met Anderson, T., Diatloff, G. and Garraty, K. fish can ingest several times its own body through the development of an adequate (1996). Potable water quality improved weight per day of submerged vegetation. and easily used detection system for sub- by harvesting the weed cabomba. 5th In Arkansas, fish stocked at 22 fish ha-1 merged aquatic weeds. The SAVEWS sys- Queensland Weeds Symposium, gave complete cabomba control in less tem for the hydro-acoustic detection and Longreach. Weed Society of Queens- than five years. In a Florida lake, control mapping of submerged water plants be- land Inc. was achieved in two years by a residual ing developed by the Tennessee Valley Anon. (1995). Botanical Electronic News population of 84 kg ha-1 (17 fish) and the Authority and the US Army Corps of En- No. 96, 27 March 1995. URL:gopher:// only changes attributable to grass carp gineers (Sabol and Melton 1995) is worthy freenet.victoria.bc.ca:70/11/environ- were an increase in nitrate-nitrite, pre- of consideration for the easy detection of ment/Botany/ben. sumably due to the decomposition of fae- cabomba infestations in impoundments. Balciunas, J.K. and Burrows, D.W. (1996). cal plant material from the fish. Native Integrating the regular use of such a moni- Distribution, abundance and field host- fish populations did change, but with no toring system into the routine manage- range of Hydrellia balciunasi Bock discernible pattern in relation to carp ment of reservoirs and impoundments (Diptera: Ephydridae) a biological con- populations (Beach et al. 1978). would go a long way to meeting the sec- trol agent for the aquatic weed Hydrilla In South Carolina, sterile (triploid) ond constraint through enabling early verticillata (L.f.) Royle. Australian Jour- grass carp are being used to control control efforts. nal of Entomology 35, 125-30. Hydrilla and Elodea (de Kozlowski 1991). Control practices need to be integrated Beach, M.L., Lazor, R.L. and Burkhalter, Sterility is ensured through fish being into the general management of the im- A.P. (1978). Some aspects of the envi- checked by three different facilities. Carp poundment, canal or river. Since cabomba ronmental impact of the white amur with a minimum length of 25 cm are is likely to be able to establish in farm (Ctenopharyngodon idella (Val.)) in stocked at a rate of 60 per vegetated hec- dams, landholders also need to be aware Florida, and its use for aquatic weed tare. of this potential and integrate checks for control. Proceedings of the 4th Interna- Whilst grass carp appear to be an effec- the weed into their general property man- tional Symposium on the Biological tive biological control agent for aquatic agement plan. Control of Weeds, Gainesville, USA, weeds such as cabomba and their effect on The control practices used must be tai- pp. 269-89. native ecosystems can be reduced by us- lored to the particular type of water body Buckingham, G.R. and Bennett, C.A. ing sterile triploids, their release into Aus- being treated. In a potable water supply, (1989). Laboratory host range of tralian waters is not likely to be acceptable very regular mechanical harvesting may Paraponyx diminutalis (Lepidoptera: due to their general herbivorous habit, be the only viable method. If an impound- Pyralidae) an Asian moth adventive in their pest potential, and infestations of ment can be taken off-line, then a suitably Florida and Panama on Hydrilla cabomba being insufficient to warrant timed drawdown and a chemical treat- verticillata (Hydrocharitaceae). Environ- consideration of carp as a control agent. ment of the root mass may be an available mental Entomology 18, 526-30. Invertebrate herbivores of cabomba are option. If drawdown is not an available Carr, G.W., Yugovic, J.V. and Robinson, poorly known; the larva of the moth option, the infestation may be thinned by K.E. (1992). ‘Environmental weed inva- Paraponyx diminutalis Snellen attacks an initial chemical treatment and the re- sions in Victoria’. (Department of Con- cabomba (Buckingham and Bennett 1989) maining plants removed by hand. If im- servation and Environment and Eco- but also attacks a wide range of other poundments flow, or could overflow, into logical Horticulture Pty. Ltd., Mel- aquatic plants and is probably unsuitable catchment headwaters, containment plans bourne). as a biological control agent. Adults of the must be put into place which will stop Chittenden, F.J. (ed.) (1951). ‘Dictionary of larval leaf-mining fly Hydrellia balciunasi cabomba washing into the river system. gardening’, Volume 1. (Clarendon Bock have been recorded from cabomba in Press, Oxford). Queensland (Balciunas and Burrows Acknowledgments Clugston, J.P. (1990). Exotic animals and 1996). The polyphagous snail Marisa The Center for Aquatic Plants, University plants in aquaculture. Reviews in cornuarietis (L.) was reported as feeding on of Florida, kindly undertook a biblio- Aquatic Science 2, 481-9. an unidentified cabomba species in labo- graphic search of their Aquatic Plant In- Dawson, F.H. (1989). Ecology and man- ratory tests in Puerto Rico by Ferguson formation Retrieval System, for which we agement of water plants in lowland and Butler (1966). are most grateful. We thank Dr. John streams. In ‘Freshwater Biological As- Revilla et al. (1991) reported the pres- Madsen for providing a copy of his paper sociation Fifty-Seventh Annual Re- ence of the free-living nematodes on cabomba. Much of the information on port’, pp. 43-60. (Freshwater Biological Dorylaimus spp., Rhabditis spp. and cabomba in Queensland and its manage- Association, Ambleside, UK). Mononchus spp. on aquarium collections ment was obtained during conversations de Kozlowski, S.J. (1991). Lake Marion of cabomba from Malaysia, but these gen- with T. Anderson and G. Diatloff (Alan sterile grass carp stocking project. era are commonly found on aquatic Fletcher Research Station, Department of Aquatics 13, 13-16. plants. Natural Resources), K. Garraty (Noosa Diatloff, G. and Anderson, T. (1995). Shire Council), P. Bell and R. Rainbird Chemical control of cabomba. In ‘Tech- Management strategies (Caloundra Shire Council) and we thank nical Highlights 1994/5: Reports on There are two major problems constrain- them for their time and consideration. In- weed and pest animal control research ing action in relation to cabomba. Firstly, formation on the commercial use of conducted by the Land Protection new infestations are difficult to detect cabomba was provided by A. Birkill, Branch, Queensland Department of since inspections for this type of weed are spokesperson for the pet industry Joint Lands’, ed. M. Hannan-Jones. (Queens- not regularly made and, as a fully sub- Advisory Council on Aquatic Plants. land Department of Lands, Brisbane, merged aquatic plant, it is not easy to see Queensland). 164 Plant Protection Quarterly Vol.12(4) 1997 Elakovich, S.D. and Wooten. J.W. (1989). Wales plants. Telopea 2, 705-14. Rataj, K. and Horeman, T.J. (1977). Allelopathic potential of sixteen aquatic Lapham, V.T. (1966). The effect of ‘Aquarium plants – their identification, and wetland plants. Journal of Aquatic endothol-silvex on aquatic plants in cultivation and ecology’. (TFH Publica- Plant Management 27, 78-84. Louisiana. Abstracts of the 1966 Meet- tions, Neptune City, New Jersey). Fassett, N. C. (1953). A monograph of ing of the Weed Science Society of Revilla, E.P., Sastroutomo, S.S. and Rahim, Cabomba. Castanea 18, 116-28. America, pp. 97-8. M.A.A. (1991). Survey on aquarium Ferguson, F.F. and Butler, J.M. (1966). Leslie, A.J. (1985). A literature review of plants of quarantine importance and Ecology of Marisa and its potential as an Cabomba. Report, Bureau of Aquatic their associated nematodes. BIOTROP agent for the elimination of aquatic Plant Research and Control. (Florida Special Publication No. 40, pp. 205-15. weeds in Puerto Rico. Proceedings of Department of Natural Resources, (BIOTROP, Bogor, Indonesia). the 19th Southern Weeds Conference, Tallahassee, Florida). Rice, E.L. (1979). Allelopathy – an update. pp. 454-5. Mabberley, D.J. (1987). ‘The plant book. A Botanical Review 45, 15-109. Gangstad, E.O. (1986). Chapter 9: Herbi- portable dictionary of the higher Riemer, D.N. (1965). The effect of pH, cidal, environmental and health effects plants’. (Cambridge University Press, aeration, calcium and osmotic pressure of 2,4-D. In ‘Freshwater Vegetation Cambridge). on the growth of fanwort (Cabomba Management’, pp. 223-54. (Thomas Madsen, J.D. (1996). Cabomba caroliniana caroliniana Gray). Proceedings of the Publications, Fresno, Ca). information request. US Army Corps of 19th North-east Weed Control Confer- Garraty, K., Anderson, T. and Diatloff, G. Engineers, Waterways Experiment Sta- ence, pp. 460-7. (1996). Mechanical harvesting of tion, Lewisville Aquatic Ecosystem Re- Riemer, D.N. and Ilnicki, R.D. (1968). Re- cabomba weed. Unpublished report, search Facility, Texas, 2 pp. production and overwintering of Noosa Shire Council. Manning, J.H. and Sanders, D.R. (1975). cabomba in New Jersey. Weed Science Godfrey, R.K. and Wooten, J.W. (1981). Effects of water fluctuation on vegeta- 16, 101-2. ‘Aquatic and wetland plants of south- tion in Black Lake, Louisiana. Hyacinth Riemer, D.N. and Trout, J.R. (1980). Effects eastern United States. Dicotyledons’. Control Journal 13, 17-21. of low concentrations of terbutryn on (University of Georgia Press, Athens). Martin, D.F. and Wain, R.P. (1991). The Myriophyllum and Cabomba. Journal of Gracia, W.H. (1966). The need for aquatic cabomba color problem. Aquatics 13, 17. Aquatic Plant Management 18, 6-9. weed control in Puerto Rico. Proceed- McComas, S. (1994). Aquatic weed con- Sabol, B. and Melton, R.E. (1995). Devel- ings of the 19th Southern Weeds Con- trol. In ‘Lake Smarts: The First Lake opment of an automated system for de- ference, pp. 454-5. Maintenance Handbook’, ed. S. tection and mapping of submersed Gregory, P.E. and Sanders, D.R. (1974). McComas, pp. 5-58. (Terrene Institute, aquatic vegetation with hydroacoustic Some aspects of the life history and US EPA, Office of Water, Office of and global positioning system tech- ecology of cabomba. Abstracts of the Wetlands, Oceans and Watersheds. As- nologies, Report I: The submersed 1974 meeting of the Weed Science Soci- sessment Watershed Protection Divi- aquatic vegetation early warning sys- ety of America, p. 32. sion, Washington DC). tem (SAVEWS) – System description Guerra, L.V. (1974). The use of granulated Moore, B. (1991). No one wants a fish kill: and user’s guide (Version 1.0). Joint herbicides in potable water for the con- fish can live when using Hydrothol 191 Agency Guntersville Project Aquatic trol of submerged aquatic vegetation. in weed and algae control. Aquatics 13, Plant Management. Abstracts of the 1974 Meeting of the 16-17. Sainty, G.R. and Jacobs, S.W.L. (1981). Weed Science Society of America, p. 40. Moseley, M.F., Mehta, I.J., Williamson, ‘Water plants of New South Wales’. Hanlon, C. (1990). A Florida native – P.S. and Kosakai, H. (1984). Morpho- (Water Resources Commission, Syd- cabomba (fanwort). Aquatics 12, 4-6. logical studies of the Nymphaeaceae ney, New South Wales). Hearne, J.S. (1966). The Panama Canal’s (sensu lato) XIII. Contributions to the Sainty, G.R. and Jacobs, S.W.L. (1984). aquatic plant problem. Hyacinth Control vegetative and floral structure of ‘Water plants in Australia’, 3rd edition. Journal 5, 1-5. Cabomba. American Journal of Botany 71, (Sainty and Associates, Darlinghurst, Hellquist, C.B. and Crow, G.E. (1984). 902-24. New South Wales). Aquatic vascular plants of New Oki, Y. (1992). Integrated management of Sanders, D.R. (1979). The ecology of England: Part 7. Cabombaceae, aquatic weeds in Japan: current status Cabomba caroliniana. In ‘Weed Control Nymphaeaceae, Nelumbonaceae and and prospect for improvement. In ‘Bio- Methods for Public Health Applica- Ceratophyllaceae. In ‘Station Bulletin logical Control and Integrated Manage- tions’, ed. E.O. Gangstad, pp. 133-46. 527’. (New Hampshire Agricultural Ex- ment of Paddy and Aquatic Weeds in (CRC Press, Boca Raton, Florida). periment Station, University of New Asia’. Proceedings of the International Sanders, D.R. and Mangrum, D.O. (1973). Hampshire, Durham, New Hamp- Symposium on Biological Control and Competition between Cabomba and shire). Integrated Management of Paddy and Anacharis in Black Lake, Louisiana. Pro- Hiltibran, R.C. (1974). Cabomba control: Aquatic Weeds in Asia, 20–23 October ceedings of the 26th Annual Meeting of Where do we go from here? Abstracts 1992, Tsukuba, Japan, pp. 197-213. the Southern Weed Science Society, pp. of the 1974 Meeting of the Weed Science Ørgaard, M. (1991). The genus Cabomba 361-6. Society of America, p. 41. (Cabombaceae) – a taxonomic study. Schneider, E.L. and Jeter, J.M. (1981). The Hiltibran, R.C. (1977). Cabomba control: Nordic Journal of Botany 11, 179-203. floral biology of Cabomba caroliniana We are not there yet. Abstracts of the Osborn, J.M., Taylor, T.N. and Schneider, Gray. Proceedings of the 13th Interna- 1977 Meeting of the Weed Science Soci- E.L. (1991). Pollen morphology and ul- tional Botanical Congress, 21–28 Au- ety of America, p. 63. trastructure of the Cabombaceae: corre- gust 1981, Sydney (Abstracts), p. 132. Ito, M. (1986). Studies in the floral lations with pollination biology. Ameri- Schneider, E.L. and Jeter, J.M. (1982). morphology and anatomy of can Journal of Botany 78, 1367-78. Morphological studies of the Nymphaeales. III. Floral anatomy of Parsons, W.T. and Cuthbertson, E.G. Nymphaeaceae. XII. The floral biology Brasenia schreberi Gmel. and Cabomba (1992). ‘Noxious Weeds of Australia’. of Cabomba caroliniana. American Journal caroliniana A. Gray. Botanical Magazine, (Inkata Press, Melbourne). of Botany 69, 1410-9. Tokyo 99, 169-84. Raciborski, M. (1894). Die Morphologie Sculthorpe, C.D. (1967). ‘The biology of Jacobs, S.W.L. and Lapinuro, L. (1986). Al- der Cabombeen und Nymphaeaceen. aquatic vascular plants’. (Edward terations to the census of New South Flora, Jena 78, 244-79; 79, 92-108. Arnold Ltd., London). Plant Protection Quarterly Vol.12(4) 1997 165 Skarratt, D.B., Sutherst, R.W. and Maywald, G.F. (1995). ‘CLIMEX for Windows Ver- sion 1.0: User’s Guide’. (CSIRO/CTPM, Brisbane). Stockwell, D.R.B. (1996a). The role of biological and environmental data in modelling landscape patterns. URL: http:// calf.symbiotik.com.au/Symbiotik/Doc/ Process/process.html (10 May 1996). Stockwell, D.R.B. (1996b). The GARP model. URL: http://calf.symbiotik.com. au/Symbiotik/Model/GARP/form. html (10 May 1996). Tarver, D.P. (1985). Status report on Sonar as an aquatic herbicide. Abstracts of the 38th Annual Meeting of the Southern Weed Science Society, p. 394. Tarver, D.P. (1987). Sonar: EPA approved! Aquatics 8, 25-6. Tarver, D.P., Rodgers, J.A., Mahler, M.J. and Lazor, R.L. (1978). ‘Aquatic and wetland plants of Florida’. (Bureau of Aquatic Plant Research and Control, Florida Department of Natural Re- sources, Tallahassee, Florida). Tarver, D.P. and Sanders, D.R. (1977). Se- lected life cycle features of fanwort. Jour- nal of Aquatic Plant Management 15, 18-22. Wain, R.P., Haller, W.T. and Martin, D.F. (1983). Genetic relationships between three forms of Cabomba. Journal of Aquatic Plant Management 21, 96-8. Wain, R.P., Haller, W.T. and Martin, D.F. (1985). Isozymes in studies of aquatic plants. Journal of Aquatic Plant Manage- ment 23, 42-5. Watson, L. and Dallwitz, M.J. (1995 onwards). ‘The families of flowering plants: descriptions and illustrations’. (URL http://muse.bio.cornell.edu/ delta/).