How specific is specific enough? - case studies of three rust species under evaluation for weed biological control in Australia.
Host specificity is one, if not the most critical attribute for any biological control agent to be considered for introduction into a new environment. Nevertheless, specificity is not always an absolute measure. The acceptability of any potential non-target effects will depend on a number of factors such as the extent of the non-target damage, the status of the affected species and the inherent characteristics of the receiving ecosystem; and, any decision for introduction will have to be based upon an encompassing risk-benefit analysis. The concept of "acceptable levels of host specificity" is illustrated using the case studies of three rust species which are currently under evaluation as potential classical biological control (CBC) agents for Australia: Puccinia lantanae Farlow from Peru for lantana, Lantana camara Linnaeus; Phakopsora jatrophicola Cummins from Mexico for bellyache bush, Jatropha gossypifolia Linnaeus; and Ravenelia acaciae-arabicae Mundkur & Thirumalachar from India for prickly acacia, Acacia nilotica ssp. indica (Linnaeus) Wildenow ex Delile. Based on the research conducted to date, the risks associated with the potential release of each individual pathogen as a CBC agent for Australia and, in the case of P. lantanae also for New Zealand and South Africa, are assessed. Finally, comparisons are drawn with the risk evaluation undertaken for two other rust pathogens, Maravalia cryptostegiae (Cummins) Y. Ono and Puccinia xanthii Schweinitz var. parthenii-hysterophorae Seier, H.C. Evans & Á. Romero, previously introduced for weed control into Australia.