Institut fur Mikrobiologie, TU Braunschweig, D-38106 Braunschweig, Germany

Background and objectives
To date, research on endophytic fungi has concentrated primarily on isolation and identification of endophytes and correlation of these isolates with ecological parameters and the hosts. Only few studies have dealt with the physiology of the fungal-host interactions and the colonization of the host tissue by the endophytes. There are also only a very limited number of investigations into the use of endophytic fimgi for biocontrol [1]. There are two ways of using endophytes as biocontrol agents: (1) as weak pathogens which cause disease of undesired plants, or (2) as mutualists to protect their host plants against attacks by pests or pathogens. We have checked the first hypothesis using data from our experiments with endophytic fungi and used published results to check the latter theory.

Results and conclusions
If the endophytic fungi are to function as weak pathogens they should be necrotrophic. In this context it is interesting that of all fungal isolates (5000) tested in our screening for biologically active secondary metabolises, as a group, the endophytic fungi had the highest proportion of isolates that produced herbicially active substances. This fact reinforced us to check for their possible use as mycoherbicides. We have taken such an approach in an attempt to control two serious weeds, Cyperus rotundus from semi-arid regions and Cirsium arvense from temperate climates, with their respective endophytic fungi. In contrast to our expectations and despite massive inoculations, the endophytic infections - in keeping with the strategy of these fungal isolates - remained either asymptomatic or only small necroses developed. Further preliminary investigations with dual infections of some endophytes and specific pathogens under glasshouse conditions suggest that this approach might be more promising for achieving biocontrol of undesired plants.

In the second case the mutualistic relationships of endophytes with their hosts are utilized. Diplodina acerina is such an endophyte which could be isolated from the lesions around the eggs of gall midge of maple leaves. In this case it appears that the endophyte destroys the plant cells around the gall and, finally, the gall drops out. Another example is the 'sterile red fungus' which grows in the root-cortex of wheat and causes lysis of the hyphae of the take-all fungus [2]. In both examples, the host is protected by the endophytic infection. Nevertheless, we should bear in mind that the endophytic fimgus also takes energy by consuming metabolises of its host.

If endophytic fungi are to be suitable as biocontrol agents the question arises whether or not they are influenced by integrated pest management practices such as spraying of pesticides. There are examples in which fungicides and herbicides have been shown to inhibit growth of the endophytes and others in which they have been shown to have little effect on the fungus being used as a mycoherbicide

In conclusion, and in consideration of the results presented, the potential of using endophytic fungi for biocontrol will be discussed.

1. Dorworth CE, Callan BE, 1996. In: Redlin SC, Carris IM, eds. Endophytic Fungi in Grasses and Woody Plant. St Paul: APS Press. 2. Sivasithamparam, 1990. International Conference on 'Integrated Plant Disease Management for Sustainable Agriculture', New Delhi, Abstract S5B-002.