Prioritising biological control agents for release against Sporobolus pyramidalis and Sporobolus natalensis (Poaceae) in Australia

Sporobolus pyramidalis Beauv. and S. natalensis (Steud.) Th. Dur. and Schinz. (giant rat’s tail grass) (Poaceae), invade rangelands and pastures in eastern Australia, costing the livestock industry approximately AUS$ 60 million per annum in grazing losses. Mechanical and chemical control options are costly and largely ineffective. Biological control is viewed as the most promising control option, however this management strategy has largely been avoided for grasses, due to their perceived lack of suitably host-specific and damaging natural enemies. In this thesis, the prospects for using biological control against S. pyramidalis and S. natalensis in Australia was assessed, in light of these potential challenges. Climate matching models were used to identify high-priority geographic regions within the plants’ native distributions to survey for potential biological control agents. High-priority regions to perform surveys were identified by modelling the climatic suitability for S. pyramidalis and S. natalensis in sub-Saharan Africa (i.e. their potential native ranges’), and climatic compatibility with regions where biological control is intended in Australia. High-priority regions for S. pyramidalis included: (1) coastal East Africa, ranging from north-eastern South Africa to Uganda, including south-eastern DRC, (2) some parts of West Africa, including inland regions of the Ivory Coast and western Nigeria, (3) northern Angola and (4) eastern Madagascar, and for S. natalensis included: (1) eastern South Africa, (2) eastern Zimbabwe, (3) Burundi, (4) central Ethiopia and (5) central Madagascar. Prospective control agents collected from these regions have the highest probability of establishing and proliferating in Australia, if released. In surveys of the insect assemblages on S. pyramidalis and S. natalensis in the climatically-matched region of eastern South Africa fifteen insect herbivores associated with the grasses were identified. Insect feeding guild, geographic distributions, and seasonal abundances suggest that three stem-boring phytophagous wasps, Tetramesa sp. 1, Tetramesa sp. 2 and Bruchophagus sp. 1 (Hymenoptera: Eurytomidae), have potential as control agents. Species accumulation curves indicated that additional surveys in South Africa are unlikely to yield additional potential control agents. Field host-range surveys of 47 non-target grass species in South Africa showed that Tetramesa sp. 1, Tetramesa sp. 2, and Bruchophagus sp. 1, were only recorded from S. pyramidalis and S. natalensis. Integrating field host-range with phylogenetic relationships between plant species indicated that no native Australian Sporobolus species or economic crops and pastures are expected to be attacked by these wasps. All three wasp species are predicted to be suitably host-specific for release in Australia. Three other endophagous herbivores attacked non-target native African Sporobolus species that share a close phylogenetic relationship to native Australian Sporobolus species, and therefore, demonstrate considerable risk of non-target damage. These species should not be considered as potential control agents. Under native-range, open-field conditions, Tetramesa sp. 1 caused an approximately 5-fold greater reduction in plant survival and reproductive output than Tetramesa sp. 2 and Bruchophagus sp. 1. Tetramesa sp. 1 in combination with Tetramesa sp. 2 did not significantly increase the level of damage, while Bruchophagus sp. 1 may decrease the efficiency of Tetramesa sp. 1, if released in combination. Tetramesa 1 is therefore the most promising candidate agent. Prioritising potential agents using predicted efficacy allowed otherwise equally suitable prospective agents to be prioritised in a strategic manner. Prioritising which natural enemies to target as biological control agents is a complex task. Field host range and damage assessments in the native range may provide more realistic data than typical studies performed under artificial conditions in a laboratory or quarantine. Moreover, it could assist practitioners in prioritising the most suitable agent(s) at the earliest stage in the programme as possible. This study demonstrated that grasses are suitable targets for biological control as they can harbour host-specific and damaging natural enemies.