Goodall, Jeremy Marshall2016-09-132016-09-132016Goodall, Jeremy Marshall (2016) Invasion of Compuloclinium macrocephalum (Less.) DC in Highveld grassland: ecology, control and non-target impacts, University of Witwatersrand, Johannesburg, <http://wiredspace.wits.ac.za/handle/10539/21031>http://hdl.handle.net/10539/21031A thesis submitted to the Faculty of Science, University of the Witwatersrand, in fulfilment of the academic requirements for the degree of Doctor of Philosophy. May 2016 JohannesburgThis thesis reveals previously unknown facts concerning the invasion, ecology and management of the perennial alien forb Campuloclinium macrocephalum (Less.) DC. (Asteraceae, pompom weed) in the grassland biome of South Africa. All these areas of research are of critical importance to identify the causes of invasion and prescribe best management practises aimed at reducing the density and spread of the weed and restoring these ancient grasslands. Grassland biomes throughout the world are receiving international attention because of their vulnerability to transformation, a history of ignorance regarding their evolution and prejudice as evidenced by their exploitation. There are many theories as to why alien plants become invasive outside of their native range but most are controversial, except those that support the absence of natural enemies as the primary reason for invasiveness. Few studies have attempted to empirically measure environmental and ecological factors that facilitate invasion, not to dispute the Absence of Predators Hypothesis (APH); APH regulates invasiveness but the external factors affecting habitat vulnerability to invasion remain unchanged. Based on the assessments of 80 invaded grasslands in Gauteng Province prior to the release of any host-specific biocontrol agents, it was concluded that C. macrocephalum favours disturbed grasslands. Numerous agencies of disturbance were identified; the most important being heavy grazing, abandonment (e.g. old lands) and modification (e.g. draining of wetlands). Herbivory by generalist insects was insignificant. The main drivers of native species composition in the invaded study sites were rainfall, topography and soil texture. The weed was most problematic in grasslands with a basal cover of <19% and in poor condition from a pastoral perspective. Other alien and native invasive species were also found in grasslands with C. macrocephalum. High fire frequency also appeared to exacerbate weed density. The Novel Weapons Hypothesis postulates that some invasive species transform vegetation for establishment, densification and expansion because they are allelopathic. Stems and leaves of C. macrocephalum have both rigid multicellular hairs and glandular trichomes that exude an unidentified substance. It has been speculated that allelopathy may be an important trait aiding its rapid expansion in invaded grasslands. The importance of allelopathy and competition was investigated under controlled conditions using Eragrostis curvula (perennial grass), E. tef, (annual grass) and Lactuca sativa (lettuce) as test species. Petri-dish studies proved that root and shoot extracts of adult C. macrocephalum plants had zero inhibitory effect on the seed germination in all three test species. Stunting of radicles was evident in treatments comprising leaf extracts at 10 and 25% w/v; with E. tef displaying a higher tolerance than E. curvula. Eragrostis curvula, because it was the most sensitive of the test species, was used in a pot study together with C. macrocephalum to evaluate allelopathy and interference. The biomass and growth of E. curvula was not affected by C. macrocephalum at densities of one or five plants per pot. The weed on the other hand incurred density-dependant trade-offs in size, biomass and mortality. In a separate pot study, the incorporation of weed residues into the potting medium had no impact on the growth of E. curvula. The process went one step further by analysing the association between E. curvula and C. macrocephalum from the 80 grassland assessments. Eragrostis curvula had a narrower ecological niche and was only found in disturbed grasslands on well drained soils. Campuloclinium macrocephalum invaded a broad gradient of soil types including poorly drained wetland soils not amenable for E. curvula. Competitive exclusion between the two species was not apparent. Pot studies and field observations support a degree of tolerance between C. macrocephalum and E. curvula that lends support to coexistence at a range of weed densities. The conclusions from this study were (a) C. macrocephalum is not allelopathic and allelo-chemistry cannot be inferred as a causal mechanism for the weed’s invasiveness, and (b) C. macrocephalum and E. curvula have different limiting resource requirements that enables coexistence in areas where the latter can grow. In 2006 C. macrocephalum infestations in Gauteng Province were severely damaged by Puccinia eupatorii Dietel (Pucciniaceae), a biotrophic rust identical to the P. eupatorii strain in quarantine that was imported into South Africa from Argentina as a potential biocontrol agent. An investigation of the rust’s impact on C. macrocephalum was carried out in the 80 rangeland infestations previously assessed and how pathogen pressure affected the weed’s realised niche. The rust caused premature senescence of the stems in late summer with compensatory regrowth in autumn. In contrast disease-free plants senesced in late autumn; the rootstocks remained dormant throughout winter and did not produce compensatory regrowth. No significant changes in weed density were detected and C. macrocephalum retained its realised niche. The weed remains adapted to the rust despite a 40-year separation from the pathogen. Registered herbicides should provide 80% control of the targeted plants, providing label directions concerning concentration, mixing, application and environmental considerations are adhered to. Herbicide trials were conducted at two contrasting sites in Gauteng, a wetland and a rocky grassland, to test the efficacy of picloram and metsulfuron-methyl on the control of C. macrocephalum. Both sites had dense infestations of C. macrocephalum at the beginning of the study in 2005. Herbicides were applied in either February (summer) or April (autumn) annually for three consecutive years. Monitoring continued for an additional three years after spraying was terminated. Puccinia eupatorii established at both study sites from the second year of study. Both sites also experienced wildfires and drought. As these uncontrolled factors were not anticipated in the design their individual effects could not be factored out. Mean mortality however was <80% expected of registered herbicides which can only be attributed to uncontrolled factors acting as natural constraints. The efficacy of picloram was not significantly different (P<0.05) to that of metsulfuron methyl within a season. Autumn applications are not recommended because rust damage on the leaves is too advanced to ensure adequate herbicide uptake and translocation. Herbicide persistence in the topsoil was not detected by gas chromatography. Three applications of herbicide were inadequate to bring C. macrocephalum under control in plots. It is estimated that five to seven years of herbicide application are needed to reduce the weed density to <1 plant per plot (25 m2). This effectively renders chemical control in medium to dense infestations uneconomic. Chemical control of C. macrocephalum will only be effective if there is a commitment to follow-ups and remedial vegetation management practices. The herbicide trials also looked at the impact of picloram, metsulfuron methyl and hoeing on native species and other alien species that occurred in plots. Three plant functional groups were analysed, namely native grasses, native forbs and alien broadleaf species. The herbicides were applied as broadcast sprays over three years and therefore their effect on all broadleaf species was non-selective. Hoeing twice a year for three years only targeted C. macrocephalum. High forb richness coupled with low abundance and patchy distribution meant there was a high turnover between replicate plots across the study sites. Herbicides had a significant impact on the native forb functional group compared to hoeing which had a superficial effect. Hoeing did not suppress C. macrocephalum. The parasitic forb Thesium utile was killed in all treatments except the untreated control. The target-specific hoe treatment also eliminated T. utile, suggesting a new association might exist between parasite and C. macrocephalum. Hoeing also facilitated the establishment of alien annual weeds. In South Africa the grassland biome is under considerable threat from mining, afforestation, agriculture and urban development. Although alien plants pose a minor threat in contrast to these land-transforming activities they remain the greatest threat to grasslands that are not threatened by exploitation. Campuloclinium macrocephalum is not a pioneer species, but rather a long-lived perennial herb with evolutionary adaptations to fire, herbivory and disease. These characteristics have enabled it to establish and spread in grasslands that have been maintained in a variety of disturbed states by a range of agencies that reduce grass basal cover and exacerbate soil erosion. Addressing poor land use practises that foster C. macrocephalum invasion and spread are as important as the remedial activities necessary to control the species. Grassland restoration/rehabilitation practises have not been investigated in the context of C. macrocephalum management and is of critical importance to the integrated control of the weed. Although comprehensive literature about C. macrocephalum have recently been published gaps in our understanding of its biology, ecology and control still exist that prevent the development of best management practices. These areas for new research make ideal projects for post-graduate students. Future research should focus on (a) rust–herbicide interactions, (b) fire and seedbank dynamics, (c) integrated weed management incorporating biological control, fire, selective herbicide application techniques (e.g. spot-spraying) and ecosystem restoration practises. Grazing strategies promoting grass species adapted to frequent non-selective defoliation restrict pompom weed better than degraded underutilised rangelands, as evidenced by low levels of C. macrocephalum in communal areas, and warrants further investigation, including fence-line contrasts in invaded and un-invaded road reserves.Online resource [142 pages]enCampuloclinium macrocephalumGrassland ecology--South AfricaThesis