In vitro observations of net community production (NCP) imply that the oligotrophic subtropical gyres of the open ocean are net heterotrophic; in situ observations, in contrast, consistently imply that they are net autotrophic. At least one approach must be returning an incorrect answer. We find that (a) no bias in in situ oxygen-based production estimates would give false-positive (net autotrophy) rates, (b) observed (13)C enrichment of surface water dissolved inorganic carbon (DIC) can be explained only by positive NCP (net autotrophy), (c) lateral and vertical inputs of organic carbon are insufficient to sustain net heterotrophy, and (d) atmospheric input of organic material is too small to support in vitro rates of net heterotrophy and would yield δ(13)C depletion of surface DIC, quite the opposite of what is observed in the subtropical gyres. We conclude that the in vitro observations, implying net heterotrophy, must contain a bias that is due to an underestimate of photosynthetic rate and/or an overestimate of respiration rate.