Nine different accessions of Sesbania sesban were examined for their toxicity to ciliate protozoa from sheep receiving a mixed grass hay–barley/fishmeal concentrate in Aberdeen, Scotland. Several accessions were much more toxic than those (ILRI 10865, ILRI 15036) identified and tested in a previous study. Two highly toxic accessions and one less toxic accession were compared with ILRI 15036 in a feeding trial in Debre Zeit, Ethiopia with sheep receiving a sululta hay–wheat bran diet. No suppression of protozoal numbers occurred in response to S. sesban supplementation (200 g/day). When rumen fluid removed from sheep on the supplemented diet was tested in vitro, protozoal activity appeared to be inhibited less by S. sesban in the mixed population than in either washed protozoa from the same rumen fluid or than had occurred in Scottish sheep. It was concluded that dietary, microbial or animal factors in the Ethiopian sheep had caused the destruction of the anti-protozoal material present in even the most potent accessions of S. sesban.