Although de‐airing procedures are commonly performed during cardiac surgery, use of these procedures is not necessarily based on evidence. Uncertainly remains around the size of bubbles that can be detected by echocardiography, whether embolized air or carbon dioxide can be absorbed, and the reasons for embolic events occurring despite extensive de‐airing. Since air bubbles are invisible in the blood, we used simple experimental models employing water and 10% dextran solution to determine the correlation between actual bubble size and the depicted size on echocardiography, bubble size and floatation velocity, and the absorption of carbon dioxide under embolization and irrigation conditions. Bubbles depicted as larger than 1 mm were overestimated by echocardiography: the actual size was larger than 0.4 mm in diameter. While bubbles of 0.5 mm had a floatation velocity of 2 to 3 cm/sec, the buoyancy of bubbles smaller than 0.3 mm was negligible. Thus, bubbles that are depicted as larger than 1 mm on echocardiography or that present with apparent buoyancy should be visible and need to be meticulously removed. However, echocardiography cannot distinguish bubbles of around 0.1 mm in diameter from those of capillary size (< 10 μm). Thus, we advise continuous venting of dense bubbles until they become sparse. While carbon dioxide was rapidly absorbed when circulating, the absorption of embolized carbon dioxide was negligible. These results suggest that detected intracardiac air represents residual «air,» with carbon dioxide already absorbed. Therefore, the use of conventional de‐airing procedures needs reconsideration: air and buoyant bubbles should be removed from the heart before they are expelled into the aorta; this requires timely and precise assessment with transesophageal echocardiography and effective collaboration between surgeons, anesthesiologists, and perfusionists.