Long-term monitoring of biological control agents in their areas of introduction is essential to assess their effectiveness. There is a need to monitor and evaluate agent dispersal and impacts so that the degree of success can be quantified or reasons for failure can be clearly understood. A fungal pathogen, Colletotrichum gloeosporioides f. sp. miconiae Killgore & L. Sugiyama (Melanconiales, Coelomycetes, Deuteromycetinae), found in Brazil in 1997 was released in the tropical oceanic island of Tahiti (Society Islands, French Polynesia, South Pacific) to control miconia, Miconia calvescens DC (Melastomataceae), a small tree native to Tropical America, which has invaded native rainforests. The plant pathogen, proven to be highly specific to miconia, causes leaf spots, defoliation and eventually death of young seedlings in laboratory conditions. Two permanent plots in Tahiti (Taravao Plateau and Lake Vaihiria) were monitored for a period of 6 years to assess the pathogen's dispersal and impacts on miconia in the wild. Leaf spots were observed approximately 30 days after inoculation. Percentage of infected plants reached 100% after 3 months, and between 90% and 99% of leaves were infected. Subsequent re-infection occurred after 3 months at Vaihiria and 18 months at Taravao. Mortality rate for monitored plants was 15% and reached 30% for seedlings less than 50 cm tall. Within 3 years, the fungus had disseminated throughout the island of Tahiti and had infected nearly all the miconia plants up to 1400 m in montane rainforests. It was also found on the neighbouring island of Moorea without any intentional release there. Leaf damage on miconia canopy trees increased from 4% to 34% with elevation in permanent plots set up between 600 and 1020 m. Our study showed that rainfall and temperature were two limiting environmental factors that affected fungal spread and disease development. Although this plant pathogenic agent is successfully established, has spread efficiently and has caused significant impacts on seedlings, additional biocontrol agents are still needed to fully control the massive invasion by miconia in the Society Islands.

CSIRO established its first overseas research laboratory on biological control at Montpellier in late 1966 to start a programme on skeleton weed, Chondrilla juncea L.). The laboratory was set up to develop the science to underpin effective biological control, by parallel studies in the native and introduced range of Australia's pests. Since establishment within a French research agency (CNRS), the facility moved in 1994 from rented facilities into a purpose-built CSIRO-owned facility, with support from Australian industry bodies and the French government. This facility has been CSIRO's largest long-term overseas investment in research. The core focus on biological control of weeds has been increasingly supplemented by other research activities that are not otherwise possible within Australia. We present an economic and scientific review of the laboratory on its 40th anniversary. The facility cost on average Aus$1.3 million (2006 $$) per year (67% on direct research activities and 33% of infrastructure and administration) and generated at least $27 benefit for Australia for every $1 invested. Staff produced 279 publications of which 159 are in journals that are currently ISI rated (average citation rate in 2007 was 14.8 per ISI journal paper).

Diclidophlebia smithi Burckhardt, Morais and Picanço (Hemiptera: Psyllidae) is a monophagous species which was selected as possible agent of biological control of miconia, Miconia calvescens DC. (Melastomataceae), a native plant of Central and South America that has become an aggressive invader of forest ecosystems in French Polynesia, Hawaii and Australia. The objective of this work was to study the biology and population dynamics of D. smithi in Viçosa and Dionísio (state of Minas Gerais, Brazil) from June 2001 to June 2002 and from February 2004 to February 2005 and evaluate injuries caused to the host plant and occurrence of natural enemies of this psyllid. Frequency distribution of the distance between the antennae indicated the existence of five nymphal instars. Colonies of the psyllid were observed throughout the year in Viçosa and Dionísio. Population peaks occurred from April to July (winter: a period of low temperatures, drought and short photoperiod in this region). Nymphs and adults were observed attacking buds, inflorescences and infrutescences of M. calvescens and causing damage by sucking the plant sap and injecting toxins. Desirable traits such as high population growth rate, easy mass rearing, occurrence throughout the year, host specificity, attack to reproductive organs and potential capacity to adapt to different climatic conditions including those similar to where M. calvescens invasions are occurring, all indicate that D. smithi is a promising biological control agent of M. calvescens.

Biological control efforts against the perennial invasive Euphorbia esula/virgata in North America have left 30-50% of all treated sites without impact after 10-15 years. Those efforts focused almost exclusively on insect releases. Much evidence is available indicating that soil biotic factors affect both invasiveness and biocontrol effectiveness. The authors have shown that soilborne bacteria and fungi are linked to biomass reductions or mortality in conjunction with insect damage. To understand factors possibly affecting synergistic interaction of the insects with plant pathogens shown to cause rapid weed mortality, predominant bacteria associated with the flea beetle Aphthona flava Guill. (Coleoptera: Chrysomelidae) released to control E. esula/virgata L. in western North America, were isolated and identified. Two Euphorbia-infested sites with widely differing levels of impact 8-10 years after insect release were sampled. From the site that exhibited rapid, sweeping declines in Euphorbia density, 6 of 12 isolates were Bacillus spp., 4 were coryneform species and 2 were Pseudomonad aceae. Bacteria isolated from the Cottonwood site included some species often associated with the biocontrol of soilborne plant pathogens. The results of tests for a range of hydrolytic enzymes showed that the two groups differed in the frequency of isolates positive for such enzymes as cellulase and xylanase. Two isolates from each location representative of predominant bacterial species and their range of traits were selected for testing on E. esula/virgata in combination with Aphthona spp. After 35-37 weeks, two isolates positive for cellulase from the Knutson Creek site caused significant (P = 0.05) dry weight reductions of E. esula/virgata plants of 64% and 6%, respectively, in combination with Aphthona spp. One of the two isolates from the Cottonwood site, also positive for cellulase production, caused a 60% reduction in dry weight compared with the control.

Tradescantia fluminensis Vell., also known as wandering Jew, is an herbaceous monocot native to South America. It is an invasive plant in New Zealand and the southeastern United States where it is considered highly invasive by the Florida Exotic Pest Plant Council. The pathobiota of T. flumi nensis in Brazil is almost unknown and could include phytopathogenic microorganisms that could be used in classical biological control programs. A survey for specialized, coevolved phytopathogenic microorganisms of T. fluminensis was initiated in 2003. Five fungal species have been collected including three basidiomycetes-a rust fungus (Uredo sp.), Kordyana tradescantiae (Pat.) Racib. and Ceratobasidium sp.; a hyphomycete-Cercospora apii Fresen. and an ascomycete-Mycosphaerella sp. A bacterial disease was also observed and the bacterium identified as Burkholderia andropogonis (Smith, 1911), based on morphological, biochemical and molecular methods. Its pathogenicity to T. fluminensis was confirmed, and a host-range test was performed. Unfortunately, results indicated that the bacterium is not sufficiently host-specific for classical introductions. Observations of the damage caused by fungal pathogens in the field suggest that those with the best potential as biological control agents are Uredo sp., K. tradescantiae and Mycosphaerella sp.