An initial investigation into biological control options for Schinus terebinthifolia in South Africa

Schinus terebinthifolia Raddi (Anacardiaceae) (Brazilian pepper tree) is a native tree to subtropical South America that was introduced into South Africa as an ornamental plant. Globally, it is regarded as one of the world’s worst invasive trees. In South Africa, this aggressive pioneer species is becoming increasingly problematic and is being considered as a target for biological control. In South Africa the tree has acquired a native seed-feeding wasp, Megastigmus transvaalensis Hussey (Hymenoptera: Torymidae). The wasp’s native hosts are indigenous Rhus species (Anacardiaceae), but it has expanded its host range to form a new association with both S. terebinthifolia and its close relative S. molle L. (Anacardiaceae). In order to quantify the seed predation by M. transvaalensis on S. terebinthifolia seeds, tree populations were surveyed across the Eastern Cape and KwaZulu-Natal provinces. The wasp was present at 99% of the S. terebinthifolia populations with an average of 22% of the seeds being destroyed. In the Eastern Cape Province, the highest seed damage occurred at the start of the winter months, when about 35% of seeds were damaged. This fell to less than 12% in spring and summer when the plants were flowering. Megastigmus transvaalensis may have slowed the rate of spread of the plant, but it is unlikely to reduce population sizes of S. terebinthifolia in South Africa in the long-term. Biological control efforts can be assisted by knowing the origin and invasion history of the target species. Genetic analyses are often the only way to elucidate the invasion history of invasive alien plants because it is rare to find detailed records of plant introductions. Both microsatellite and chloroplast DNA analysis were conducted on S. terebinthifolia trees from the plant’s introduced distribution in South Africa and both Florida and Hawaii, USA. These samples were compared to plants from the native distribution of South America. The analysis indicated that the S. terebinthifolia in South Africa was most likely sourced from the state of Rio de Janeiro in Brazil, which is the same source of the invasive populations in Florida and Hawaii. Importantly, the South African populations were all found to be “haplotype A”. Plants samples collected from Hawaii USA were the closest match to the South African plants. Biological control agents known to damage haplotype A which have been considered for use in Hawaii and Florida should therefore be prioritised for South Africa. Schinus terebinthifolia has a broad distribution in South Africa; however, the majority of the current distribution is limited to the coastal regions along the eastern coast in KwaZulu-Natal Province. This suggests that the species may be climatically limited. Species distribution models in MaxEnt were used to predict the suitable ecological niche of the species. Using occurrence localities from both the native and invaded range to calibrate the models resulted in 56% of the modelled areas being considered suitable for the growth of S. terebinthifolia in South Africa. This included areas in the Eastern Cape, Western Cape and Limpopo provinces. When the models were calibrated using just the native range data, or just the invaded range data, predicted distributions were more restricted and limited to the coastal areas of the Eastern Cape and KwaZulu-Natal provinces. The coastal areas between Florianopolis and Santos in Brazil were highlighted as the most climatically similar to the invasive populations of S. terebinthifolia in South Africa. These areas should be prioritised if native range surveys for potential biological control agents are conducted in South America. Although the native seed-feeding wasp is damaging to S. terebinthifolia in South Africa, the tree is still not under suitable levels of biological control and is likely to spread and increase in density. New biological control agents are therefore required. Genetic and climatic matching has determined where the most appropriate region to collect new potential biological control agents is. The genetic matching data has also indicated that biological control agents that have been released, or are being considered for release, in Hawaii and Florida, are likely to be suitable for the South African plants because they have been shown to be damaging to ‘haplotype A’. These agents should therefore be the first to be considered for release in South Africa