A rebreathing method was developed for measuring diffusing lung capacity for carbon monoxide (DLCO) in a hyperbaric environment. Twenty two professional naval divers with normal lung function were included in the study. Significant correlations were found between rebreathing and single breath measurements for DLCO (r = 0.94; p less than 0.001; standard error of the estimate (SEE) = 0.66), alveolar volume (VA) (r = 0.79; p less than 0.005; SEE = 0.51), and DLCO/VA (r = 0.83; p less than 0.001; SEE = 0.11). In 17 divers, rebreathing DLCO (DLCOrb) was also measured at 20 minutes pre-dive, during the first decompression stop of the dive to 45 m for 25 minutes, and at 10 minutes post-dive. Compressed air diving was performed in a dry walk-in chamber and the United States Navy decompression table was followed. The pressure induced decrease in the rate of CO binding to haemoglobin was adjusted to normobaric conditions using a theoretical approach. Also, the presence of venous bubbles post-dive was detected by precordial doppler monitoring. A biphasic change in DLCO was noted: initially, DLCO was increased during the dive (p less than 0.005); this was followed by a post-dive decrease; DLCO/VA changed in a similar manner, as VA was only slightly altered. Only a small post-dive precordial doppler bubble grade was found. In conclusion, rebreathing DLCO measurement is a useful respiratory function test in the hyperbaric environment. It appears that an increase in D(L)CO during the compressed air dive is related predominantly to increased pulmonary capillary blood volume caused by increased negativity of the pleural pressure, hyperoxic pulmonary vasodilatation, and cardiorespiratory centralisation of the blood. The decrease in D(L)CO post-dive was only partially related to the presence of the venous bubbles detectable by doppler.