Das Projekt "Biologische Bekämpfung der Blattschneiderameise Acromyrmex landolti mit Futtergräsern im Chaco Paraguays" wird vom Umweltbundesamt gefördert und von Universität Hohenheim, Fakultät III Agrarwissenschaften I, Institut für Pflanzenproduktion und Agrarökologie in den Tropen und Subtropen, Fachgebiet Biodiversität und Landrehabilitation in den Tropen und Subtropen durchgeführt. A major pasture pest in South America is the leaf-cutting ant Acromyrmex landolti. The objective of this study was to identify forage grasses with resistance to the attack of A. landolti by inhibiting the fungus growth. An experiment took place at the Estación Experimental Chaco Central in Paraguay. Six grasses were tested: Cynodon nlemfuensis, Panicum maximum cv. Gatton, Cenchrus ciliaris, Brachiaria brizantha, Digitaria milanjiana cv. Jarra and Urochloa mosambicensis. The experiment started on 10 March 1997 and was carried out for seven weeks. Results showed that nests were deeper and the number of chambers was larger than stated in available literature. The results indicated a higher activity at nests supplied with D. milanjiana and C. plectostachyus. With an average of more than 2/3 filled, C. nlemfuensis and D. milanjiana had the largest amount of fungus in their chambers. In both cases B. brizantha and C. ciliaris fed nests showed a significantly lower vitality than D. milanjiana and C. nlemfuensis. Dry matter consumption differed significantly between the treatments after 4 weeks of trial. From week 5 on D. milanjiana and C. nlemfuensis were significantly more consumed than all other treatments. Ants fed with B. brizantha consumed the least amount of dry matter. In the bioassay D. milanjiana plant sap enhanced the fungal growth most whereas B. brizantha sap enhanced the fungal growth least. D. milanjiana and C. nlemfuensis can be regarded as susceptible to the attack of A. landolti. B. brizantha can be considered resistant while the other grasses tested cannot be assigned to either group. The bioassay proved to be a secure, simple and quick way of testing the ability of a species to inhibit fungal growth. Further research is needed to find out the basic behaviour patterns of A. landolti.