Characterising the herbivore community and its impact on Sonchus oleraceus (Asterales: Asteraceae) in its invaded range in Australia

Sonchus oleraceus is an annual species native to Eurasia and Northern Africa that has become a major weed of cultivated fields and ruderal areas in Australia. Populations are difficult to manage in cropping systems because of the development of resistance to common herbicides. Biological control is being investigated as an additional tactic for managing the weed. A literature review was conducted to identify the phytophagous arthropod species already associated with the species in Australia to guide such a biological control programme. To identify opportunities for biological control agents to aid in management, we undertook field surveys across Queensland and South Australia in different environments. We also investigated factors that may influence their performance in Australia. Both the literature review and field surveys identified 21 arthropod species associated with S. oleraceus in Australia, most of them being generalist species and pests of exotic origin. Capitula were the most damaged plant part while stems were relatively free from insects, except aphids. The field surveys recorded an unexpected new interaction between S. oleraceus and the gall midge, Contarinia jongi (Diptera: Cecidomyiidae). This association was surprising as the midge, only known to develop on Alstroemeria (Liliales: Alstroemeriaceae), a very distant relative to S. oleraceus (Asterales: Asteraceae), was reported in Australia only a few years ago under greenhouse conditions. The midge and the moth Eublemma cochylioides (Lepidoptera: Erebidae) were the two species that occurred most frequently in developing flower head samples. We considered their infestation rate as a proxy of herbivory and tested whether the environment surrounding the plant may influence herbivory. Both E. cochylioides and C. jongi showed the highest infestations in ruderal sites compared with the sites located in conventional farming areas. We discuss the implications of our results in the context of selecting and releasing candidates for the biological control of the weed, especially in agricultural landscapes.