A DNA probe specific for the V and VI transmembrane domains of the bovine oxytocin receptor was initially prepared by reverse transcription PCR, and its structure and specificity confirmed by DNA sequencing. This probe was then used to screen a bovine genomic DNA library in bacteriophage lambda, and three positive clones were purified, subjected to restriction analysis and relevant fragments sequenced. Parallel to this, a cDNA library prepared using bovine endometrial RNA at the time of ovulation was screened by PCR employing the same primers as above. The longest cDNA clone was also fully sequenced. This clone still lacked, however, a substantial stretch of 5'sequence. The full transcript structure, and hence also the exon-intron organization, was then obtained by RT-PCR using primer oligonucleotides deduced from the cloned genomic sequence. All nucleotide sequence information was derived from a combination of two independent genomic clones, a cDNA clone and several independent RT-PCR reactions programmed by myometrial RNA, all in both strand orientations. The structural organization of the bovine oxytocin gene essentially conforms to that described for the human gene. Unlike the human gene, however, the 5'non-coding region of the primary transcript is interrupted by only a single intron, with a further intron in the coding region separating the sequences encoding the transmembrane domains VI and VII. The difference between this structure and that for the human gene suggests the existence of a differential splicing of 5' non-coding sequences.