An existing cDNA library of the segmented double-stranded RNA genome of maize rough dwarf virus, a plantinfecting reovirus of the genus Fifivirus was used to further characterise three genome segments. Segments $7, $8 and S10 were completely sequenced and found to consist of 1936, 1900 and 1802 base pairs, respectively. They all contained the conserved structural motifs present in all genome segments of plant-infecting reoviruses. Computer analysis indicated that the coding strands of segments $7 and S10 each contained a single large open reading frame [ORF], consisting of 591 and 558 codons, respectively, as is the norm for reovirus genome segments. Segment $8 contained two large non overlapping ORFs consisting of 347 and 209 codons, located in the 5' and 3 ~ terminal domains, respectively; an arrangement already described for segment $6. In vitro translation experiments confirmed that the gene products of gel-purified dsRNA genome segments $7 and S 10 were single polypeptides of 68 and 63 kDa, respectively. Transcription/translation of a plasmid carrying a full-length cDNA copy of segment $8 yielded two (40 and 25 kDa) polypeptides. The nucleotide sequences of segments $7 and S 10 were aligned with those of the corresponding segments of rice black-streaked dwarf virus, another member of the genus Fijivirus, and found to be almost identical.