1. The formation of bubbles is the basis for injury to divers after decompression, a condition known as decompression illness. In the present study we investigated the effect of endurance training in the rat on decompression-induced bubble formation. 2. A total of 52 adult female Sprague-Dawley rats (300-370 g) were randomly assigned to one of two experimental groups: training or sedentary control. Trained rats exercised on a treadmill for 1.5 h per day for 1 day, or for 2 or 6 weeks (5 days per week) at exercise intervals that alternated between 8 min at 85-90% of maximal oxygen uptake (VO2,max) and 2 min at 50-60% of VO2,max. Rats were compressed (simulated dive) in a decompression chamber in pairs, one sedentary and one trained, at a rate of 200 kPa x min(-1) to a pressure of 700 kPa, and maintained for 45 min breathing air. At the end of the exposure period, rats were decompressed linearly to the 'surface' (100 kPa) at a rate of 50 kPa x min(-1). Immediately after reaching the 'surface' (100 kPa) the animals were anaesthetized and the right ventricle was insonated using Doppler ultrasound. 3. Intensity-controlled interval training significantly increased VO2,max by 12 and 60% after 2 and 6 weeks, respectively. At 6 weeks, left and right ventricular weights were 14 and 17 % higher, respectively, in trained compared to control rats. No effect of training was observed on skeletal muscle weight. Bubble formation was significantly reduced in trained rats after both 2 and 6 weeks. However, the same effect was seen after a single bout of aerobic exercise lasting 1.5 h on the day prior to decompression. All of the rats that exercised for 1.5 h and 2 weeks, and most of those that trained for 6 weeks, survived the protocol, whereas most sedentary rats died within 60 min post-decompression. 4. This study shows that aerobic exercise protects rats from severe decompression and death. This may be a result of less bubbling in the trained animals. The data showed that the increase in aerobic capacity per se was not the main mechanism, but rather an acute effect that was most notable 20 h after a single, or the last, exercise bout, with less effect after 48 h.