Most Mentha species have 1-25% 1-limonene and 0.5-8% 1,8-cineole, but 19 individuals having more than 50% limonene-cineole were found in a progeny of 10,000 Strain 2 M. citrata - M. crispa F 1 hybrids. When the same strain of M. citrata (2 n = 96) having the genotype I1I1 i 2 i 2, a lavender herbage odor with oil assaying 30% linalool and 58.5% linalyl acetate, is hybridized with the closely related octoploid species M. aquatica (2 n = 96) having the genotype i 1 i 1 i 2 i 2, a menthofuran herbage odor with oil assaying 65-80% menthofuran, the fertile F 1 hybrids should have the genotype I 1 i 1 i 2 i 2 and a lavender odor with oil assaying 84-90% linalool - linalyl acetate. In addition to 111 normal lavender-odored hybrids, this cross gave one individual (Strain 38) having 20.4% limonene and 36.4% cineole and one individual (Strain 625) having 67.5% limonene and 23.6% cineole. Since M. aquatica is homozygous for menthofuran production, and since Strain 38 - M. aquatica backcross progenies had the disomic ratio of 1 limonene and cineole-odored: 1 methofuran-odored, it is evident that the 57% limonene - cineole content of Strain 38 is due to a single dominant gene Lm. Strains 38 and 625 were hybridized with other tester species having known genotypes for other oil constituents to demonstrate that the gene Lm prevents the conversion of limonene to more advanced compounds, namely: carvone, pulegone, methofuran, menthone, menthol, and menthyl acetate which are normally developed in the oil of other species having the recessive gene lm. Strain 38 hybrids with M. citrata show that the dominant I gene interrupts oil biogenesis at an earlier stage than the Lm gene and largely prevents the synthesis of limonene and cineole. Nine of 21 strains having 57- 94% limonene - cineole were investigated. Strains 38 and 62 had the genotype i 1 i 1 i 2 i 2 Lm 1 lm 1 lm 2 lm 2 or i 1 i 1 i 2 i 2 lm 1 lm 1 Lm 2 lm 2, whereas Strain 625 and six others had the genotype i 1 i 1 i 2 i 2 Lm 1 lm 1 Lm 2 1m 2. These segregants from the segmental allopolyploids may be explained by assuming that M. aquatica has the genotype [Formula: see text] and Strain 2 of M. citrata the genotype [Formula: see text] with A and a designating the non-homologous centromere regions of the two chromosome pairs carrying the linked genes on different chromosome arms. Crossing over between the genes would not be detectible when there is normal autosyndetic bivalent pairing, whereas occasional quadrivalent pairing of the four chromosomes of Strain 2 of M. citrata could lead to gene interchanges between chromosomes non-homologous for the centromere region.