In deep-sea diving the respiratory gases are forced into solution in the body fluids and tissues as a result of increased pressure. When return to the surface is fairly rapid, these gases are present in a supersaturated state, and bubbles frequently form. Extremely small at first, these go through a period of expansion, followed by gradual absorption. Nitrogen gas, being physiologically inert, is most important in causing these bubbles; its inertness results in their slow absorption. Ascent to altitudes of 25,000 feet or higher in aircraft results in a similar supersaturation of dissolved gases, with a tendency to bubble formation. The intense pain of “caisson disease,” commonly spoken of as diver's “bends,” has its counterpart in aviator's “bends,” resulting from exposure to extremely high altitude. The symptoms can readily be elicited by simulated flight in an altitude chamber. Ordinarily the pain disappears at about 25,000 feet during the return to ground-level pressure. It is thought that the “bends” pains are due to expanded extravascular bubbles, as a direct or indirect pressure effect. It has also been suggested that intravascular bubbles may, through ischemia, stimulate the pains of “bends.” Symptoms are more frequent and more severe if exercise is engaged in at the high altitude. They are most commonly referred to the joints (1). With the intent to clarify the etiology of aviator's “bends,” we have made a roentgen study of the affected parts in a decompression chamber at altitudes of 35,000 and 38,000 feet, both in the presence and absence of pain. As our work progressed, we learned of work being done by others in the same field and have received many helpful ideas from them. The first reported roentgen observation of gas in the tissues in “bends” seems to have been by Gordon and Heacock in 1940 (2). A patient who had been working under pressure (25 lb. per sq. inch) suffered fracture of the proximal ends of the tibiae and was “decompressed” rapidly. Study of the published reproductions of the roentgenograms suggests that the authors may have been misled by a lipohemarthrosis. This is not an uncommon occurrence after fracture at this site, and a layer of fat floating on top of blood in the synovial cavity can indeed imitate a bubble of gas. Methods In our studies lateral views of knees were made on non-screen film with a small portable x-ray machine. Often pain was present in only one knee, but films of both knees were made (routinely) for control. Similar films were also made of the subjects at sea level. No extensive studies of the chest or abdomen were attempted. Results In every person, whether with or without “bends” pain, gas can be shown in the knee joint at an altitude of about 20,000 feet, increasing to a maximum in about thirty minutes at 30,000 feet. The demonstration of large volumes of gas (estimated at 50 to 75 c.c.) is clearly seen in Figure 1.