Here we report that petal tissue was responsible for the bulk of the phenylpropanoid emission. The activity of S-adenosyl-i-methionine:(iso)eugenol O-methyltransferase (IEMT), a nove1 enzyme that catalyzes the methylation of the para-4'-hydroxyl of both eugenol and (iso)eugenol to methyleugenol and isomethyleugenol, respectively, was also highest i n petal tissue. IEMT activity was absent from floral tissues of plants not emitting (iso)methyleugenol. A C. breweri cDNA clone encoding IEMT was isolated, and its sequence was shown t o have 70% identity t o S-adenosyl-L-methionine:caffeic acid Omethyltransferase. l h e protein encoded by this cDNA can use eugenol and isoeugenol as substrates, but not caffeic acid. Steadystate IEMT mRNA levels were positively correlated with levels of IEMT activity in the tissues, and no IEMT mRNA was observed in flowers that do not emit (iso)methyleugenol. Overall, the data show that the floral emission of (iso)methyleugenol is controlled at the site of emission, that a positive correlation exists between volatile emission and IEMT activity, and that control of the level of IEMT activity is exerted at a pretranslational step.Flowers of many plants attract pollinators by producing and emitting a complex mixture of low-molecular-weight volatile compounds. Floral scents may function as both long-and short-distance attractants and nectar guides to a variety of animal pollinators (for review, see Dobson, 1993). Since insects are able to distinguish between complex floral scent mixtures, discriminatory visitation based on floral scent has important implications for population structure and reproductive isolation (Dodson et al