Within Australia, the Sporobolus indiens complex (Gramineae) consists of 22 entities. Two of these entities - S. indicus var major (Giant Parramatta Grass GPG) and S. pyramidalis (Giant Ratstail Grass GRTG) - are serious weeds of pasture in coastal areas of northern New South Wales and Queensland. A third entity, S. jacquemontii (American Ratstail Grass ARTG), may also be considered a weed in some areas. Recent increases in the area infested have led to serious economic loss to agricultural production, notably the livestock industry. A combination of the grasses' invasiveness and low palatability to stock leads to rapid degradation of marginal pasture often rendering it useless. The potential to use endemic, ovary invading pathogenic fungi to control weedy members of the S. indicus complex was explored and assessed in this thesis.A series of preliminary experiments developed techniques which allowed accurate quantitative assessment of disease severity, a requirement of later experiments. A qclearing and stainingq technique, frequently used in the study of foliar biotrophs was modified. This technique cleared previously opaque floral structures while staining invasive fungal tissue within the ovary. This allowed rapid quantitative assessment of disease severity. The tissue most susceptible to infection was found to be anthesing florets and this tissue type was the subject of all further experiments. For optimum disease development an inoculated host was subjected to two periods of high humidity during which there was free water on the leaf surface broken by a 24 to 36 hour period of low relative humidity.The three weed species and a number of related native species were subjected to an analysis of molecular genetic variation. Random Amplified Polymorphic DNA (RAPD) technology was chosen to develop and compare profiles of intraspecific and interspecific variation for a number of entities. This provided resolution of the taxonomic uncertainty. The implication of genetic variation to biological control systems was also considered. This technique was able to distinguish plants at the level of genotype and showed greater variation amongst samples of S. pyramidalis than S. indiens var major. Greater genetic variation may be indicative of a lower amenability to control through biological means.Four species within the genus Bipolaris were assessed for their pathogenicity toward inflorescences of five Sporobolus entities. Both B. ravenelii and B. Crustacea caused severe disease on all entities. B. ravenelii infected an average of 71% of S. indiens var major ovaries and 62% of S. pyramidalis ovaries. There were no significant differences between levels of infection caused by B. crustacea, which averaged 71%. This study provided quantitative confirmation of the studies of Alcorn (1982).A field study was conducted at the New South Wales Department of Agriculture's Grafton field station during the weed's growing season (November- June) during 1993-94. The objectives of the study were to identify key correlates tondisease spread in the field, examine the effect of infestation density upon disease incidence and severity and to examine plant behaviour following infection. Mean daily temperature one, two and three days before the day on which samples were assessed for disease correlated strongly with disease incidence. The regression equation derived for this relationship showed that the overall trend was for temperature to fall during this period but a degree of fluctuation was present from day to day. The regression equation also indicated that increasing variation between nightly and daily leaf wetness periods may increase disease incidence.n n n n n n n n n n n