Colletotrichum gloeosporioides is the causal agent of anthracnose disease of mangoes. Infection occurs when humidity is high and rain-dispersed spores germinate and form an appressorium on immature mangoes. The infection then becomes quiescent until the fruit is harvested. On ripe fruit infection is visible as black, sunken lesions on the surface. At the pre-harvest stage, the disease is controlled with the application of a range of fungicides, and at the post-harvest stage by hot benomyl treatment. The extensive use of benomyl, both pre- and post-harvest, has resulted in the occurrence of isolates of C. gloeosporioides resistant to this fungicide. To devise an alternative strategy of disease control, the potential for biological control of anthracnose has been investigated. Potential microbial antagonists of C. gloeosporioides were isolated from blossom, leaves and fruit of mango, and screened using a series of assay techniques. In total 650 microorganisms, including bacteria, yeasts and filamentous fungi, were isolated and tested for their inhibition of growth of C. gloeosporioides on malt extract agar. Of these 650 isolates, 121 inhibited the fungus and were further tested on their ability to inhibit spore germination in vitro. Of these, 45 isolates, all bacteria and yeasts, were inoculated onto mangoes, which were artificially inoculated with C. gloeosporioides, and assessed for their potential to reduce the development of anthracnose lesions. A further selection was made, and 7 isolates were chosen to be used in a semi-commercial trial in the Philippines. This final screening procedure yielded two potential candidates for field trials, isolate 204 (identified as Bacillus cereus) and isolate 558 (identified as Pseudomonas fiuorescens). A field trial involving pre-harvest application of the biological control agent, was conducted using isolate 558. This isolate was chosen for this purpose since in in vitro experiments it significantly reduced germination of C. gloeosporioides spores. In the field trial 558 was applied in combination with nutrients and compared to treatments which had received no treatment or which had received conventional fungicide (benomyl) application. On spraying, high numbers of 558 were recorded on the leaf surface, but no reduction in post-harvest development of disease was observed. Failure of disease control was attributed to rapid death of the bacterium on the phylloplane. Inpost-harvest trials, isolates 204 and 558 were both tested in combination with different application methods, including the addition of sticker, peptone, fruit wax or a sucrose polyester. Application of 204 did not reduce disease development. Application of 558, however, did significantly reduce anthracnose development compared to the control fruit. No additional benefit was achieved by incorporating the bacteria in peptone, fruit wax or sucrose polyester. The mode of action of isolate 558 was investigated in detail. There was no evidence for parasitism taking place, or the production of volatile compounds, in the suppression of disease development. No antibiotic compounds were detected, but isolate 558 did produce a siderophore. A sharp increase in pH was also observed in culture media in which 558 was grown. Disease control may result from a combination of these two factors.