The complete sequence of the 2193 residues of RNA 3 (Mr 746000) of cucumber mosaic virus (Q strain) was determined by a combination of chemical and enzymic sequencing techniques utilizing cloned DNA fragments. The nucleotide sequence of RNA 3 also gave the complete sequence of 1027 residues of RNA 4 (Mr 349000), which codes for the viral coat protein and is derived from the 3' end; there are 53 untranslated nucleotides at the 5' end of RNA 4. The nucleotide sequence provided the amino acid sequences of the two proteins coded for by RNA 3: the 5'-terminal 3A protein with 333 amino acids (Mr 36700) and the 3'-terminal viral coat protein with 236 amino acids (Mr 26200). These two coding regions are in the same reading frame, are separated by an intercistronic region of 123 nucleotides and are flanked by two untranslated regions of 94 nucleotides at the 5' terminus and of 263 nucleotides at the 3' terminus. Secondary structure models are postulated for parts of the RNA 3 sequence. These are considered to be important in the control of the translation and replication of RNA 3 and in the processing of RNA 3 to give RNA 4.