The deterioration of essential oils and products containing essential oils has been a serious problem in the food flavor and essential oil fields for many years. A study of terpene autoxidation would give a more thorough understanding of the reactions occurring between the terpene, oxygen, and the catalyst. Although these reactions are known to be quite complex (2, 4, 5. 6, 7, 13, 14, 15, 16, 18, 19), involving initiation of a free radical chain reaction, chain propagation with oxygen participation, and chain termination, a systematic study of several of the more important variables would help clarify some of the more cryptic aspects of the over-all problem. The natural choice for a substrate is limonene since this terpene is one of the most widely distributed and abundant hydrocarbons found in the plant kingdom. It occurs in numerous volatile oils and is present in some as the main constituent, especially in the citrus oils. n-Limonene has been identified in oil of orange (about 90% ), lemon, mandarin, lime, grapefruit, bergamot, neroli, petitgrain, elemi, caraway (400/o), dill, fennel, celery (60%)) erigeron, and orthodon (8).EXPERIMENTAL The course of the autoxidation reactions was followed by oxygen uptake which was determined by conventional techniques using a Warburg respirometer (17). The main compartment of the reaction flask contained 0.2 ml (1.23 x lo4 moles) of n-limonene unless otherwise noted, and the center well of the vessel contained 0.2 ml of distilled water. It was found to be extremely difficult to obtain accurate and meaningful readings of the manometers when the oxidations were carried out in the absence of water. This is probably due to some anomalous effects of vapor pressure encountered between the water in the manometer fluid and the substrate. In later experiments a modified Erlenmeyer flask was employed to provide a greater surface area for oxidative attack, and the water was added directly to the substrate. [Similar techniques using aqueous dispersions have been reported by Borglin et nl (5)].The authors were unable to distinguish any difference between oxidations carried out in the two flasks. All autoxidation experiments were conducted in duplicate, and frequently a number of replicate determinations were carried out. A control experiment was run concurrently with all manometric examinations.The control vessel contained all the constituents of the actual reaction mixtures examined with the exception that an inert gas, nitrogen, filled the vessel instead of air.The gas-liquid chromatographic apparatus employed to separate the products of the autoxidation of n-limonene was an Aerograph mode1 A-90-C.b Columns were constructed of stainless steel tubing, g inch o.d. and 10 feet in length.The support material used throughout this study was Sil-0-Cel C-22 diatromaceous earth firebrick (30-60 mesh). Fractions of this were sieved to size and further graded by sedimentation in water, after which they were dried at 110" C for 17 hr. The liquid phase was introduced by deposition from benzene so...