Solanum sisymbriifolium (sticky nightshade) is a shrubby weed of South American origin that was introduced to South Africa at the turn of the century. Despite being indicative of disturbed habitats, the weed was found to be invasive in conservation, agricultural recreational and suburban areas; this, coupled with the failure of both chemical and mechanical control attempts suggested that the weed was a good candidate for biological control. A biological control programme which followed a standard protocol was initiated. Observations suggested that S. sisymbriifolium dispersed primarily by seeds. Plants produced large quantities of fleshy fruit, favoured by frugivorous birds, which facilitated the rapid spread of the weed into new habitats. The seeds germinated quickly, especially in disturbed soil, often below the parent plant where they dropped from burst fruit, and along fences where birds roost. The pre-introductory survey of the weed revealed that S. sisymbriifolium was attacked by a relatively small number of, mainly polyphagous, herbivorous insects. These were localised and sporadic in incidence and inflicted very little observable damage. The herbivore fauna of S. sisymbriifolium was depauperate even in relation to two other exotic weeds, S. elaeaglllfolium and S. mauritianum, in South Africa. The paucity of native herbivores on S. sisymbriifoliwn was ascribed to a combination of the weed's taxonomic distinctness from South African Solanum species, and the dense covering of glandular trichomes on its leaves. Although it was shown that the exudate produced by these glandular trichomes of S. sisymbriifolium seriously impeded the movement and feeding of native herbivores, there was not enough evidence to suggest that the glandular trichomes, alone could have been responsible the lack of herbivores on the weed. Two leaf-feeding Cassidinae Gratiana spadicea and MetJ-iona elatior were screened as agents for the biological control of S. sisymbrilfolium. Favourable biological characteristics for both species included a high rate of increase, long-lived adults, many generations per year, and a high per capita feeding rate. Host range was investigated in larval survival tests and adult choice tests. The larvae of both species were reared through to the adult stage on several of the native Solanum species tested, and also on eggplant (S. melongena). However, the survival of G. spadicea on the majority of these species was very low, suggesting that the beetles would be unlikely to attack them in the field. This was supported by the adult choice tests, where G. spadicea females displayed, a strong oviposition preference for their natural host. In contrast, M. elatior larvae survived well on non-host plants, and the females selected several non-host species, including eggplant for oviposition. It was argued that the conflict of interests involving eggplant was overrated because eggplant is subjected to a stringent insecticide spray regime. Based on this evidence, permission for release was granted for G. spadicea. The impact of native parasitoid host range extensions to weed biological control agents in South Africa was investigated. Native parasitoids were recorded from nearly half of the agent species that had established on their target weed. The level of concealment and taxon influenced susceptibility of the agents to parasitoid attack Poorly concealed endophagous agents were most susceptible to attack, while exposed feeders were fairly free from attack. However, native parasitoids were reported not to strongly influence weed biological control agent populations and it was concluded that no agent should be rejected based only on its susceptibility to native parasitoid attack. Finally, several predictions are made as to the potential success of G. spadicea on S. sisymbriifolium and some of the challenges facing the biological control of weeds are discussed.