Post release evaluation of the distribution and efficacy of Eccritotarsus catarinensis and Eccritotarsus eichhorniae on Pontederia crassipes in South Africa

Biological control involves the release of new species into the environment and therefore, needs to be carefully monitored through post-release assessments which have been largely neglected in the science. Post-release evaluations of biological control programmes reveal whether the control agent has established and if it impacts weed demography, while cost-benefit analyses require a different set of data that show the magnitude on return on investment. The biological control effort on Pontederia crassipes in South Africa uses, amongst others, two species of mirid, Eccritotarsus catarinensis and E. eichhorniae. Initially, they were released as a single species, but were recently divided using molecular techniques. Eccritotarsus catarinensis was released in 1999, and E. eichhorniae in 2007. After many releases over two decades, there was need to assess where each species was established in the country. Molecular techniques proved to be valuable in identifying the two species as they are morphologically indistinguishable in the field. Therefore, molecular techniques should be routinely used for screening biocontrol agents, whether new or as re-introductions. Annual surveys of the mirid release sites around South Africa were undertaken between 2016 and 2019. At each site both insect and plant parameters were measured. Only E. catarinensis is established in the field in South Africa despite the multiple releases of E. eichhorniae at over 70 sites across the country, and E. catarinensis has established at only 22 of the 45 release sites accessed during this study. This thesis tested climate, interaction with other agents already on P. crassipes, and direct competition between the two mirid species as reasons for the lack of establishment of E. eichhorniae. The results of the country-wide surveys showed that climate and water trophic status were the major determinants in the establishment of E. catarinensis. Most of the establishment was recorded in the warmer regions of the country, however, a few populations of the mirid also established in cooler areas, thus demonstrating a degree of thermal plasticity, and possible microclimates as the mirids persisted at sites shaded by riparian vegetation. Stochastic events such as active herbicide campaigns, winter frosts, droughts and floods were responsible for the absence of the mirid at some sites. At some of the eutrophic sites, despite the abundance of E. catarinensis, plants still proliferated as the water trophic status facilitated plant growth, thus, plants were able to compensate for the damage inflicted by the mirid. A more intensive, monthly, post-release evaluation was conducted on the Kubusi River, Eastern Cape Province between 2016 and 2019. This is regarded as one of the cooler water hyacinth sites. Populations of biological control agents at this site fluctuated seasonally. At this site, cold winters caused frosting of the leaves of P. crassipes with the exception of plants growing under overhanging vegetation that provided a refuge for the mirid. But, cool temperatures in the winter months (May to August) severely reduced the populations of E. catarinensis that required a long recovery phase in spring. The consequence of this was that the plants grew unchecked from the onset of the growing season forming dense mats. Of the four agents at the Kubusi River site, Eccritotarsus catarinensis recovered slowest after winter, with lag phases ranging from two months to several months of the three-year period. The release of a suite of agents has implications on the agents themselves, where interactions between the agents can be important. Interactions between pairs and even multiple agents can have implications for biocontrol, where agents are either complimentary or interfere with each other. In this case, because E. catarinensis recovered the slowest of the four agents at the site, plants were of a poor quality by mid-summer resulting in low mirid populations. Competition in weed biological control could be expected to be strongest between pairs of agents that share the same niche, and this could be the reason why E. eichhorniae failed to establish at sites where E. catarinensis had already been established for several years. When the two mirids were combined in manipulated trials in a polytunnel, populations were lower compared to when the two mirids occurred separately. Under warm conditions, it is likely that E. eichhorniae would be the superior agent compared to E. catarinensis. The evaluations discussed in this thesis highlighted gaps in agent release methodology in multispecies settings, as well as the need for strategic augmentation pre- and post-winter. It is important to release agents that will complement each other rather than compete, therefore, when releasing agents in a multispecies setting, niche differentiation needs to be considered. Here it is concluded that the best practice for dealing with the mirids is that they should be released individually, and at sites that have no other biological control agents in order to ultimately assess their efficacy. Landscape level, long-term monitoring of biological control programmes shows the impact of the control programme at a broader scale and, are far more informative than short-term studies and at fewer sites. Long-term post-release evaluations should be mandatory in biological control programmes. Furthermore, these assessments will help develop new strategies or improve on existing ones, thus achieve greater success in control