Some naphthalene and phenoxy compounds prevent preharvest drop of apples, pears, and citrus fruits. These studies have been complicated by an unrecognized high level of ethylene produced by leaves and fruit on trees sprayed with these growth regulators. An apparent contradiction is the effectiveness of both 2,4-dichlorophenoxyacetic acid and n-dimethylaminosuccinamic acid (a growth retardant which retards biosynthesis of auxin) in preventing abscission of apples. Thus, in the presence of low auxin concentrations in the tissue, this growth retardant prevents fruit abscission even more effectively than 2,4-dichlorophenoxyacetic acid at high auxin concentrations in the tissue. This anomaly is clarified on the basis that n-dimethylaminosuccinamic acid, in the presence of a known low ethylene biosynthesis, delays maturity of the fruit and thus prevents fruit abscission. On the other hand, 2,4-dichlorophenoxyacetic acid prevents abscission by direct growth hormone action, in spite of the side effects of ethylene production which speeds ripening of the fruit. With the promotion of abscission of leaves and fruit of agricultural crops, attention is given to the use of chemicals which induce ethylene production when applied to the plant, but which have no growth promotion effect to retard abscission. We can distinguish 5 kinds of such chemicals. One group includes gibberellic and abscisic acids that induce treated leaves to produce ethylene and abscise (under certain circumstances). However, they do not induce ethylene production by fruit and do not promote fruit abscission. A second group includes ascorbic acid, which, when used at relatively high levels, induces fruit to produce enough ethylene to promote abscission. Ascorbic acid-treated leaves also produce ethylene but not enough to cause much defoliation. A third group includes protein-synthesis inhibitors, such as cycloheximide. When low concentrations (about 30 mumoles/l) are sprayed on the fruit, the rapid effect of the freely moving ethylene (produced by the treated fruit) appears to mask temporarily any potential effect of the slowly moving inhibitor. A fourth group includes 2-chloroethylphosphonic and cupric ethylenediaminetetracetic acids, which induce ethylene production of fruit and leaves; production by leaves is substantially greater than by fruit and substantial defoliation results. A fifth group includes the cotton defoliation chemicals which clearly produce ethylene primarily as a result of chemical injury to the leaf blade. Another group of compounds, represented by beta-hydroxyethylhydrazine, produces ethylene by a chemical reaction with formaldehyde and water, and the presence of leaves or fruit is not required. At this time we are unaware of how chemicals in groups one to four act to promote ethylene evolution in leaves and fruit, but possible biological and chemical paths of ethylene production are discussed.