Researchers using the eddy covariance approach to measuring trace gas fluxes are often hoping to measure carbon dioxide and energy fluxes for ecosystem intercomparisons. This paper demonstrates a systematic microprocessor- caused lag of 20.1 to 20.2 s in a commercial sonic anemometer-analog-to-digital datapacker system operated at 10 Hz. The result of the inadvertent negative lag (i.e., the digitized analog concentration signal is received before its corresponding instantaneous wind and temperature signal) is a loss in the magnitude of the recorded measured flux. Based on raw field data specific to the system used in this study (2.6-m sample height; roughness length 5 3 cm), errors in flux measurements due to a 0.2-s lag ranged from 10% to 31%. Theoretical flux errors, based on ideal near-neutral cospectra, for a 0.2-s phase shift range from 21% to 55% for neutral-stability wind speeds of 0.5-15 m s21. The application of a 0.2-s phase correction improved an early-summer, sage shrubland ecosystem energy balance by 29.5%. Correction equations for lag times of 0.1-0.4 s at the sample height of 2.6 m are provided.