In a survey conducted in Zimbabwe in 1993, waterhyacinth was present in seven out of the eight provinces. No control measures were imposed on 35% of the infested dams and 61% of the infested rivers, while in 47% of the infested dams and 11% of the infested rivers control of waterhyacinth was being attempted with a combination of 2,4-D and mechanical control methods. The population of Neochetina eichhorniae and N. bruchi declined during the period 1993 to 1995 in the Hunyani River system. Several fungi were isolated from diseased waterhyacinth, and Fusarium moniliforme (isolate 2ex 12), F. solani (isolates 5a ex25 and 2a3), and F. pallidoroseum (isolate 3ex1) were found to be the most pathogenic. Large numbers of viable conidia were produced in shake-flask liquid fermentation with modified Richard's medium and in solid fermentation with food grains. Conidia production in straw was poor with the exception of waterhyacinth straw. Host range studies conducted in pots and in the field indicated that Commelina benghalensis was moderately susceptible to both isolates of F. solani in the field, while Setaria verticilata grown in pots was moderately susceptible to isolate 2a3. Brassica rapa and Crotalaria juncea grown in pots were moderately susceptible to F. moniliforme but they showed no infection in the field. Fifty-nine additional plant species of ecological and agricultural importance were not susceptible to the Fusarium species. When F. solani, F. pallidoroseum and Neochetina spp. were used individually in ponds, they did not control waterhyacinth. When the fungi were combined with Neochetina spp., the area covered by waterhyacinth and the volume of waterhyacinth were significantly reduced.

Phomopsis convolvulus Ormeno, a foliar pathogen of field bindweed, is a good candidate to be developed as a bioherbicide. Large numbers of infective propagules were produced in shake-flask liquid fermentation with modified Richard's (V-8) medium and in solid-substrate fermentation with pearl barley grains. In complex liquid media, pycnidium-like structures were observed. Most conidia stored at $-$70$ sp circ$C remained viable and virulent for at least six months. / In controlled environment studies, a minimum of 18 hr of dew was required for severe disease development on inoculated plants. The addition of gelatin, Sorbo $ sp{ rm TM}$, or BOND$ sp{ rm TM}$ to the inoculum did not enhance the disease under various leaf wetness periods. A continuous dew period of 18 hr was superior to the cumulative effect of three interrupted 6 hr dew periods. Secondary inoculum was produced on diseased plants placed under moist conditions for 48 hr or more. / In greenhouse experiments, seedlings at the cotyledon and 3- to 5- leaf stage were severely diseased and killed when inoculated with 10$ sp9$ conidia/m$ sp2$. This inoculum density adversely affected the regenerative ability of 4 wk old seedlings and established plants, but few plants were killed. Inoculation of the healthy regrowth from plants previously inoculated with the fungus resulted in much less disease symptoms than expected.

No description available.

Jointed cactus, 0puntia aurantiaca Lindley, remains a problem and continues to expand its range in South Africa, in spite of a mandatory herbicidal control. programme. The feasibility of biologically controlling the weed is being explored because the cost of herbicides has escalated and, if it succeeds, biological control is permanent self perpetuating and relatively cheap. This thesis describes the introduction and release in South Africa of the phycitid moth, Tucumania tapiacola Dyar, as a potential biocontrol agent against jointed cactus. A preliminary objective of the study was to resolve the taxonomic confusion within the genus Tucumania, so that the various populations of the moth from widespread localities and from different host plants could be identified. The efficiency of mass-producing T. tapiacola in the insectary was continually improved through investigations of the insects biology, and its response to various environmental parameters. Techniques were developed to manipulate the different life-stages of T. tapiacola so that every release was made with the maximum possible number of individuals, all in the same stage of development. Trials were made with various methods for transporting and releasing T. tapiacola in the field, and the most successful of these were employed during the establishment programme. In all, approximately 830 000 T. tapiacola eggs, larvae and adults have been released at seven localities in South Africa, between May 1977 and February 1982. So far, the moth has failed to establish for more than three to five generations at any release site, for reasons that were not immediately apparent. The mortality factors acting against the immature stages of T. tapiacola have been investigated and quantified. The accumulated data were used to construct partial-life-tables and survivorship curves. These show that survival of the eggs, larvae and pupae differed in shaded and exposed habitats, and on small, medium and large O. aurantiaca plants. The overall mortality suffered by the immature stages alone did not account for the establishment failure. Alternatively, genetic problems that are commonly associated with the collection, mass production and release of biocontrol agents may have been responsible for the failure. Methods of overcoming these problems during future releases are discussed.

No description available.

No description available.

No description available.

No description available.

A computer simulation model written in PL/I of the flowering phenology of C. thoermeri and the seasonal population dynamics of R. conicus was constructed based on a synthesis of the literature, and on laboratory experiments which determined developmental periods, mortality for immature stages and the ovipositional pattern of R. conicus. The model successfully simulated the ovipositional pattern of R. conicus and the flowering phenology of C. thoermeri. Larval population dynamics of R. conicus, in the model however, did not fit the validation data. Though the model was temporally synchronized with the field data, the model results were one degree of magnitude less. The poor fit was attributed to sampling errors in the validation data. Studies of the overwintering mortalities of both organisms were performed. R. conicus was found to overwinter with lowest mortality on musk thistle rosettes; mortality was 73.6% in 1981, higher than previously reported but possibly more representative of field conditions. No overwintering mortality was observed for established rosettes of C. thoermeri. Studies of seed dispersal of C. thoermeri were performed to determine the percentage of seeds removed from the vicinity of the plant where they originated. Seed dispersal observed for different wind velocities demonstrated that less than 1% of seeds are blown further than 100m and most seeds are deposited within 50m of the release site. A Gaussian model of seed dispersal based on experimentally determined parameters and literature values corroborated these results. / Master of Science

Conflicting opinions about the effectiveness of H. pakistanae as a biological control agent for hydrilla prompt researchers to find a method for assessing the fly's success. Developing a physiological age-grading system for the fly using ovarian morphology to detect changes in reproductive activity is useful for evaluating reproductive status of the fly in field populations. Changes in the appearance of follicular relics in ovaries with oviposition provide a reliable method to estimate fecundity. Characteristics of follicular relics were used to develop a system with eight physiological age classes, three nulliparous and five parous. Changes that occur in the fat body were used to assist in classification of nulliparous females or those with low egg counts.