Date of this Version
Renewed debate over the risk of non-target effects in biological control reflects, in part, the recent quantification of direct and indirect ecological effects of the flowerhead weevil, Rhinocyllus conicus Frol., in North America. To help resolve the issue, we review the published data for R. conicus from both Europe and North America: pre-release (1961-1968), post-release (1969-1985) and more recent (1986-1999). Our aim was to determine the extent to which host range expansion on to native North American species, and the associated ecological effects, were predicted or predictable. Our overall conclusion is that more was known than is generally realized, yet more information would have been required to complete the initial assessment of ecological consequences. Three important points emerge. First, the potential effect of R. conicus on native North American species was not a major element of the testing programme. Second, the host range expansion observed is consistent with the pre-release and early post-release data, and so was predictable, if not predicted. The pre- and early post-release data showed that R. conicus could feed and develop on multiple Cirsium species, including two North American species. Third, we found that the studies needed to quantify the likely magnitude of feeding by R. conicus on North American Cirsium species, and thus the ecological consequences of that feeding, were not done. Instead, inferential arguments were used to suggest that any feeding by R. conicus on North American species would not be substantial. We conclude that there were sufficient data, which suggested that North American Cirsium species would be acceptable host species, to have warranted further testing to define and quantifY the potential ecological side-effects of introducing R. conicus to North America. Contemporary concerns should now mandate such tests.