Twelfth Australian Weeds Conference

LANTANA CAMARA BIOCONTROL: CAN NEW TECHNOLOGIES HELP?

M.D. Day and M.A. Hannan-Jones Alan Fletcher Research Station, PO Box 36, Sherwood, Qld 4075

Abstract The methods used in research for the bio- causing substantial damage to the weed and the plant logical control of Lantana camara, a hybrid species is still a serious pest of many countries in Africa, Asia whose progenitors are of Caribbean origin, were ex- and the South Pacific. amined and the application of new technologies as- The main factors influencing the establishment of sessed. Though biological control of L. camara has agents on L. camara are the identity of the plant from been conducted for nearly 100 years, lantana remains which the agents were collected, the phenotype of L. a major problem in many countries. camara in the target country and climate (M.D. Day New developments and recent technologies are being and S. Neser unpubl. data). L. Smith (unpublished implemented in the biological control of L. camara. report) suggested that as a result of hybridisation, there These include genetic studies of the host plant, the use were no lantana species in Mexico or the Caribbean of modelling tools such as climatic matching and the that corresponded with the phenotypes occurring in conduct of ecological studies evaluating the agents’ Australia or many other countries. Consequently, impact on the target weed. DNA studies suggest that agents found on Lantana spp. in their native range are L. camara in Australia has a greater affinity to some not necessarily adapted to attack L. camara in Aus- Lantana species than others. In addition, a greater tralia or elsewhere. proportion of released agents have established when Agents that have been released, have shown prefer- collected from some Lantana species than from oth- ence for some lantana phenotypes over others. Haseler ers. The studies of the agents’ performance on differ- (1966) reported that the moth Neogalea sunia preferred ent lantana phenotypes in the target country suggest the pink and white L. camara phenotypes in the field, that agents may show a preference for some pheno- while the moth Salbia haemorrhoidalis preferred the types. Climate matching has assisted in the selection red lantana phenotypes. Harley (1973) reported that of suitable areas in which to release agents, while eco- Teleonemia scrupulosa also showed a preference in logical studies monitor the effectiveness of the agents. the field, being more commonly found on pink-edged Such developments have resulted in a greater appre- red phenotypes than the common pink phenotype. ciation in the complexities of plant/insect interactions, In addition, L. camara in Australia is found over a the factors that influence host acceptance and estab- wide geographic and climatic range from Cape York, lishment of agents on L. camara. It is hoped that this Qld in the north to Moruya, NSW in the south, from new information may eventually lead to more efficient sea level to 1000 m, and in areas receiving from 650 biological control. mm of rainfall per year to areas receiving over 3500 INTRODUCTION mm. As a consequence of this variability, only two agents occur in all areas infested with lantana and many Lantana camara is a hybrid species whose progeni- have only a limited distribution. tors originated in tropical America. The plant was bred in Europe as an ornamental prior to its introduction to This paper outlines possible ways in which new tech- Australia and many other countries. As a result of nologies and methods may assist in overcoming these hybridisation, there are now over 650 phenotypes limiting factors and achieving better biological con- worldwide (Howard 1969), with 29 occurring in Aus- trol of lantana. tralia (Smith and Smith 1982). Biological control of NEW DEVELOPMENTS AND TECHNOLOGIES L. camara has been conducted since 1902, when po- tential agents were first sent to Hawaii from Mexico. Taxonomic and Genetic Studies L. camara is an Since then, 35 agents have been introduced worldwide, aggregate species derived through horticultural and with approximately 60% establishing (Julien and natural hybridisation, selection and somatic mutation Griffiths 1998). However, only four species are from a number of similar and probably closely related

330 Twelfth Australian Weeds Conference tropical American species. This has resulted in many garcia and Falconia intermedia have all shown pref- hundreds of “named” forms, varieties or phenotypes erence for, and performed better on some lantana phe- (Howard 1969). These vary in ploidy, bush shape, notypes than on others with laboratory populations flower colour, prickliness, chemical composition and dying out on the lesss preferrred phenotypes (Day et response to the environment and natural enemies al. 1998; Day et al. 1999; A. Urban et al. unpublished (Swarbrick et al. 1995). data). This has led to confusion when trying to identify re- Climate has also been suggested as a factor influenc- lated forms in their native range on morphological ing the establishment, distribution and effectiveness grounds. Smith and Smith (1982) stated that of agents. Field surveys in eastern Australia have found “L. camara L. sensu stricto has not been seen from that only two agents (Ophiomyia lantanae and Australia”, however, Munir (1996) said “The type of Lantanophaga pusillidactyla) are present throughout the typical variety matches well with the Australian the entire distribution of L. camara. Others, such as material of the species”. Of the 29 phenotypes of L. Uroplata fulvopustulata and Leptobyrsa decora, have camara that Smith and Smith (1982) described from very limited distributions, being found only in far north eastern Australia, only four resembled taxa that L. Queensland. Smith collected during his survey in Central and South Leaf feeding insects, which have proved successful in America in 1966/67. The two most common pheno- the control of many weeds, have had limited impact types in Australia (“Common Pink” and “Common on L. camara in many regions. Though agents such as Pink-Edged Red”) were not sighted by Smith during Octotoma scabripennis and Uroplata girardi may sea- his survey. sonally defoliate lantana, reducing its vigour and caus- There has also been some debate of the affinity of L. ing a reduction in flowering, the overall biomass of camara to other lantana species based on morphologi- the weed is not necessarily affected, and plants do not cal grounds. L. camara appears to share common char- die. Leaf feeders and flower and seed feeders have acters with L. urticifolia, L. hispida, L. moritziana and proved less effective in subcoastal areas where L. L. hodgei (Lopez-Palacios 1977, Sanders 1987, Munir camara can often become yellow, lose its leaves and 1996). stop flowering during the cool, dry winter months. Consequently, these agents are not able to sustain suf- Recent attempts have been made to place the Austral- ficiently high populations over winter to control L. ian material into context with the New World species camara. using molecular biology techniques. A genetic study (L. Scott unpublished data) of Australian L. camara Modelling tools Potential release sites of agents and cv “Common Pink” Smith and Smith showed that this the agents’ subsequent population growth can be pre- important widespread weedy phenotype has a close dicted using a number of software packages, such as affinity with the Mexican material known as CLIMEX and DYMEX, which have been developed L. urticifolia. The study also showed this phenotype by Sutherst and Maywald (1985, unpubl. data). is close to L. tiliifolia from Brazil, which has been CLIMEX is a computer model for matching climates. considered as a subspecies of L. urticifolia (R. Sand- The “Compare Locations” function uses a set of pa- ers unpubl. data). rameters to describe a species response to the climate The significance of these studies is that a greater per- of a given location. The matches of this routine are cent of agents collected from L. urticifolia or based on comparisons of long-term climate averages L. tiliifolia have established in Australia, than have of the species’ native range and that of the areas of its potential agents collected from other Lantana species. intended release. DYMEX models the life cycle and potential population growth of an insect given inputs Further genetic studies are needed to clarify the posi- such as climate, the agent’s biology and parasitism tion of the other Australian phenotypes, and weedy rates. lantanas in other countries in terms of their relation- ship to the New World Lantana species. These programs can be used to predict the suitability of release sites, compare various collection sites of Ecological Studies Laboratory studies confirm the biological control agents with target site, and the in- field observations that show agents may have a prefer- teraction of the life cycle of the agent with the target ence for one or more phenotypes of lantana. weed. Kriticos (1997) used these software packages Aconophora compressa, Charidotis pygmaea, Ectaga to model the predicted climatic range of Acacia nilotica

331 Twelfth Australian Weeds Conference and the most susceptible life stage of the plant for bio- progress and impact on the weed in each area. Pre- logical control. dicted population growth can then be determined us- ing models such as DYMEX. However, this program DISCUSSION is still being fully developed. Meanwhile data on the The use of new technologies such as genetic studies agents’ biology is being collected for future use. and modelling programs, combined with ecological While it is hoped that the implementation of these tech- studies, is expected to provide the necessary informa- niques will improve the establishment rate of agents, tion to make better decisions in the research into the it is accepted that the availability of host specific agents biological control of L. camara. Results of genetic may be limiting. studies suggest that some New World species of lantana, particularly L. urticifolia, are more closely ACKNOWLEDGEMENTS related to the Australian L. camara than others. This The authors wish to thank their colleagues at the De- is supported by having a greater percent of agents that partment of Natural Resources for their helpful com- were collected from L. urticifolia establish on L. ments on the manuscript. This paper was part of a camara than those collected from other lantana spe- collaboratory project between the Alan Fletcher Re- cies. Therefore, potential agents should be collected search Station and Cooperative Research Centre for from the lantana species most closely related to L. Tropical Pest Management. camara in the target country. REFERENCES Recent surveys in Mexico and the Caribbean have found a number of insect species on L. urticifolia that Day, M.D., Willson, B.W. and Latimer, K.A. (1998). show promise as biocontrol agents. These include the The life history and host range of Ectaga garcia, a cerambycid, Parevander xanthomelas and a root feed- biological control agent for Lantana camara and ing chrysomelid beetle, Longitarsus sp. L. montevidensis in Australia. BioControl 43, 325-338. Phenotype preference studies conducted in the labora- tory confirm that some agents prefer certain pheno- Day, M.D., Willson, B.W. and Nahrung, H.F. (1999). types to others. These agents are now being reared The life history and host range of the golden lantana and released on the most suitable phenotype and in the beetle, Charidotis pygmaea (Col.: Chrysomelidae), case of A. compressa, have established in the field on a biological control agent for Lantana montevidensis these phenotypes. Where this agent was released on (Verbenaceae). Biocontrol Science and Technol- other phenotypes, establishment had not occurred. ogy 9, (in press). Consequently agents should be reared and released on the most suitable and preferred phenotypes. Harley, K.L.S. (1973). Biological control of Lantana in Australia. In ‘Proceedings of 3rd International In addition, ecological studies suggest that in areas Symposium on Biological Control of Weeds,’ eds where L. camara drops its leaves in winter months, A.J. Wapshere, pp. 23-29. (Commonwealth Insti- agents that attack stems or have a diapause stage would tute of Biological Control, Farnham Royal, UK). be the most suitable agents for release. As a result of these studies two agents, the diapausing chrysomelid, Haseler, W.H. (1966). The status of insects introduced Alagoasa parana and the stem boring beetle, for the biological control of weeds in Queensland. Aerenicopsis championi have been released in more Journal of the Entomological Society of Queens- appropriate areas. However, it is too early to confirm land 5, 1-4. their establishment. Previous attempts to achieve es- Howard, R.A. (1969). A check list of cultivar names tablishment of these agents have failed. Therefore the used in the genus lantana. Arnolida 29, 73-109. biology and behaviour of agents should also be con- sidered when selecting release sites. Julien, M.H. and Griffiths, M.W. (1998). Biological Control of Weeds: A World Catalogue of Agents CLIMEX has been used for a number of agents in- and their Target Weeds. 4th Edition, pp89-108 cluding A. compressa and A. championi in assisting in (CAB International, Wallingford, UK). the selection of possible release sites. Subsequent monitoring has been conducted to assess the agents’

332 Twelfth Australian Weeds Conference

Kriticos, D. (1996) A Climate Analysis of the Poten- Scott, L.J., Graham, G.C., Hannan-Jones, M. and Yeates. tial Distribution of Prickly Acacia in Australia. A D.K. (1997). DNA profiling resolves the report prepared for the Queensland Department of limited importance of flower colour in defining Natural Resources, Land Protection Branch, Bris- varieties of Lantana camara. Electrophoresis 18, bane. 1560-1563. Lopéz-Palacios, S. (1977). Verbenaceae “Flora de Smith, L.S. and Smith, D.A. (1982). The naturalised Venezula. In “Flora de Venezula”: 18-655. Lantana camara complex in eastern Australia. De- (Universidad des Los Andes, Merida, Venezuela). partment of Primary Industries, Botany Branch, Brisbane. Munir, A.A. (1996). A Taxonomic review of Lantana camara L. and L. montevidensis (Spreng.) Briq. Sutherst R.W. and Maywald, G.F. (1985). A compu- (Verbenaceae) in Australia. Journal of the Adelaide terised system for matching climates in ecology. Botanic Gardens 17, 1-27. Agric. Ecosystems and Environ. 13, 281-299. Sanders, R.W. (1987). A new species of lantana Swarbrick, J.T., Willson, B.W. and Hannan-Jones, M.A. (Verbenaceae) from Dominica, Lesser Antilles. (1995). The Biology of Australian Weeds 25. Journal of the Arnold Arboretum 68, 343-348. Lantana camara L. Plant Protection Quarterly 10, 82-95.

333