Phylogenetic and chemo-ecological aspects affecting the host selection behaviour of the thistle tortoise beetle Cassida rubiginosa

Most species of insects are selective feeders that not only choose the plant but also choose specific plant organs. Insect host-finding behaviour involves several cues such as olfactory, visual, tactile and gustatory. The responses to these cues depend on the context in which the signals are perceived and on the host-insect system. Knowing the process of host-finding behaviour of insects is important for management practices. The oligophagous, chrysomelid beetle Cassida rubiginosa was introduced to New Zealand in 2007 mainly to control Cirsium arvense (Californian thistle), one of the most noxious weeds found in the world. New Zealand does not have any native thistles and very few economically important ones. Therefore, there is potential for using this beetle to target secondary weeds and to better manage the beetle for weed biocontrol. Host selection behaviour of the beetle was investigated by conducting experiments in a phylogenetic context giving emphasis to chemical cues. Host range was investigated for constitutive volatiles as well as induced volatiles. Host-range testing with constitutive volatiles was conducted by using 19 Asteraceae plants (16 Cardueae and 3 non-Cardueae) which were selected according to phylogeny. A series of single-choice and dual-choice olfactometer experiments and adult host-choice experiments were carried out investigating the beetles’ choice of constitutive volatiles. Adult host-choice experiments were performed to test the feeding and oviposition preference of the beetle. Volatiles were induced by conspecific larval damage and 13 Cardueae plant species were used. Using induced volatiles, the differential attraction of C. rubiginosa in olfactometer experiments comparing undamaged leaves with conspecific larval damaged leaves was investigated. Volatiles were collected and analysed using GC-MS. Compatibility of the beetle can change according to other biocontrol agents. Investigations were carried out with two fungal pathogens (Puccinia punctiformis and Sclerotinia sclerotiorum) infesting C. arvense by studying the olfactory behaviour of adult C. rubiginosa towards fungus-infected C. arvense. A series of olfactometer experiments were conducted to evaluate the host selection of the beetle; dual-choice experiments were performed to investigate its feeding preference. Adult beetles were given the choice between healthy thistle leaves and leaves infected by one of the pathogens. Volatiles were collected from healthy and fungus-infected plants and analysed using GC-MS. The multi-targeting biocontrol potential of C. rubiginosa was tested with marsh thistle (Cirsium palustre), a species closely related to the primary host and shown to be equally preferred in previous experiments. To test if the beetle can reduce the fitness of marsh thistle a potted plant experiment was established with four treatments (0, 50, 100 and 200 larvae/plant). Plant growth (width, height and number of branches) and reproductive performance (number of flowers, seeds, seed weight and % germination) parameters were measured. The results show that as phylogenetic distance from the primary host plant increases, the beetles’ preference for alternative hosts decreases. This was the case for all three parameters: olfactory choice for constitutive volatiles, feeding and oviposition preference. Olfactory experiments showed similar host range to host-choice experiments across the Cardueae tribe. When presented with herbivore-induced volatiles (HIPVs) the beetle increased attraction towards five hosts that it was not attracted to when tested with constitutive volatiles. Chemical compounds such as cis-β-ocimene and β-caryophyllene were detected in the HIPVs blend in the five species of plants that showed beetle attraction, expanding the olfactory host range of the beetle. However, the beetles’ attraction towards C. arvense was reduced when infected with the biotrophic rust fungus P. punctiformis while the necrotroph S. sclerotiorum had no effect. Both pathogens, however, reduced adult beetle leaf consumption. Volatile profiles showed a unique blend of compounds where the P. punctiformis-infected thistle leaves were characterised by the emission of benzenoids and indole, S. sclerotiorum-infected leaves by green leaf volatiles. Thus, having multiple biocontrol agents can reduce the efficacy of at least one of the involved agents and should be taken into consideration. The multi-targeting potential of the beetle is predicted to follow the phylogenetic trend in host preference. As such, the greatest multi-targeting potential is likely for the equally preferred species, particularly the Cirsium and Carduus genera. Though the beetle was found to be equally attracted to C. palustre in all of the experiments, it did not show an effect in reducing C. palustre individual plant growth or reproductive performance under the imposed experimental conditions. If the beetle is to have an impact on other thistle weeds, it will likely need to attack early growth stages (seedlings), or smaller size classes of plants to prevent reproduction.