The isolation and characterization of cDNAs encompassing the full length of chicken, cow, rat and human elastin mRNA have led to the elucidation of the primary structure of the respective tropoelastins. Large segments of the sequence are conserved but there are also considerable variations which range in extent from relatively small alterations, such as conservative amino acid substitutions, to variation in the length of hydrophobic segments and largescale deletions and insertions. In general, smaller differences are found among mammalian tropoelastins and greater ones between chicken and mammalian tropoelastins. Although only a single elastin gene is found per haploid genome, the primary transcript is subject to considerable alternative splicing, resulting in multiple tropoelastin isoforms. Functionally distinct hydrophobic and cross-link domains of the protein are encoded in separate exons which alternate in the gene. The introns of the human gene are rich in Alu repetitive sequences, which may be the site of recombinational events, and there are also several dinucleotide repeats, which may exhibit polymorphism and, therefore, be effective genetic markers. The 5' flanking region is G+C rich and contains potential binding sites for numerous modulating factors, but no TATA box or functional CAAT box. The basic promoter is contained within a 136 bp segment and transcription is initiated at multiple sites. These findings suggest that the regulation of elastin gene expression is complex and takes place at several levels.