Measurements of short-term, local air-sea exchange of CO2 by eddy correlation in the atmosphere from surface towers have shown that the transfer (piston) velocities in coastal areas are very large in comparison to long-term oceanic estimates from radioisotope studies. The latter agree with radon evasion and laboratory investigations involving nonreactive gases. Horizontal atmospheric advection seems to be the most likely source of significant error in the eddy correlation estimates but is probably not the cause of the large transfer velocities because they were measured in a wide range of conditions by independent investigators. Furthermore, extrapolation of the large transfer velocities measured by the eddy correlation measurements to world average air-sea exchange rates does not provide a realistic basis on which to evaluate the validity of the local eddy flux measurements in coastal areas. Important chemical and physical phenomena affecting CO2 exchange rates may be quite different in coastal as opposed to open-ocean conditions, and further research is needed in both cases.