Andes virus, one of five hantaviruses known to cause hantavirus pulmonary syndrome (HPS), emerged in 1995 in southwestern Argentina (López et al. (1996) Virology 220, 223-226). The complete nucleotide sequence of Andes virus S genome segment was determined and compared with sequences of viral RNAs in autopsy tissues of more recently reported HPS cases from southwestern Argentina and south of Chile (cases ESQ H-1/96 and CH H-1/96). Andes virus S segment was found to be 1876 nucleotides in length and to encode the nucleocapsid protein (N), 428 amino acids in length. S segment analysis also revealed a long 5' non-coding region (547 nucleotides) which displays three copies of an octanucleotide sequence repeat. Comparisons of S segment sequences of ESQ H-1/96 and CH H-1/96 (82% of the entire genome sequence) with the corresponding sequences of Andes virus revealed identities of 97.2% and 98.5%, respectively. Sequence motifs identical and in the same positions as exhibited in Andes virus 5' non-coding region were found in both, ESQ H-1/96 and CH H-1/96 sequences. Three genome fragments of the M segment sequence of the viruses (representing approximately 34% of the entire sequence) were also analyzed. Comparisons of S and M segment sequences of Andes virus with the corresponding sequences of ESQ H-1/96 showed S and M segment identities which differ by less than 1.4%. Andes virus and CH H-1/96 have S segments that differ by 1.5% from one another while their M segment fragments differ by 5.5-8.2%. Phylogenetic analysis showed that Andes virus along with ESQ H-1/96 and CH H-1/96 form a distinct lineage within the clade containing Bayou and Black Creek Canal viruses. It also showed that Andes virus branch of trees derived from comparisons of S or M sequences differed. It is concluded that Andes virus variants causing HPS circulate east and west of the Andes mountains.