For investigating the effects of ethylene gas on plant material (e.g., fruit ripening or morphological responses) known concentrations in the test atmosphere of less than 1,000 parts per million (0.1 %) are usually wanted. These atmospheres have been variously obtained. Some workers have used ripe fruit or cultures of Penicilliumn d igitatum Sacc. as sources of a low (but unknown) concentration of ethylene. Others have used flowmeter combinations, bubble counting systems, or pressure cylinders made up to specified concentrations. In closed systems, low known concentrations of ethylene may be achieved by pipetting small amounts of gas into the system or (for very low concentrations) by pipetting portions of ethylene-saturated water or dilutions thereof (8). For systems in which a continuous flow of the test atmosphere is desired, accurate regulation of the trace component is difficult. Combinations of capillary flowmeters, such as have been used for modified atmosphere studies (1, 2), are not sufficiently sensitive for regulation of the trace component, although quite satisfactory for major constituents of a mixture. (Capillary flowmeters are to be preferred over rotameters, because they can be made cheaply in any laboratory for any flow rate; they also give more stable readings). If a Mariotte bottle is used to regulate the flow of the trace constituent, the desired accuracy can be achieved in a system that will remain stable and can be calibrated and adjusted easily. This method was used and described by Huelin (6), but when his thesis was published it was not described (7). Griffiths and Potter refer briefly to the technique (5). The Mariotte bottle (4, 11) should be better known to plant physiologists. Its theory has been described by Schwertz (12) in an application to flowmeter calibration, and applications in post-harvest physiology have been shown by Platenius (9) and Pratt (10). The Mariotte bottle will maintain a constant very small flow of liquid under a constant pressure; hence, it can be used to displace a small flow of gas from a reservoir through a flowmeter into a larger air flow. Adjustments of the relative rates of flow of the trace gas and the air make a wide range of concentrations readily available. Huelin's basic method (6) is described in this note with the inclusion of modifications which extend its range of usefulness. In our laboratory, the method has been used to maintain concentrations ranging from 0.003 to 1000 ppm.