The Southern Great Plains Experiment in 1997 provided an opportunity to set up a single reference eddy covariance system and net radiometer next to ten other flux-measuring systems in Oklahoma. The net radiometers were different models from one manufacturer, except for the reference unit, and most agreed within 10%; one unit differed by 20%. This suite of systems comprised seven eddy covariance systems from several manufacturers and three Bowen ratio systems from two manufacturers. Comparisons between the reference eddy covariance system and other eddy covariance latent heat flux measurements showed RMS differences of 49 W/m2, while those between the reference eddy covariance system and Bowen ratio systems showed RMS differences of 128 W/m2. Unlike the Bowen ratio system, the eddy covariance system does not assume energy budget closure. Despite the difference in manufacturers, all systems had similar closure rates of near 80% with high latent heat fluxes and 90% with low latent heat fluxes. After forcing closure, RMS differences for the comparison between the reference and other eddy covariance latent heat fluxes dropped to 27 W/m2, while those between the reference eddy covariance system and Bowen ratio systems dropped to 99 W/m2. The coefficient of variation (COV) among raw, eddy-covariance latent heat fluxes was about 25%, and the eddy covariance and Bowen-ratio systems differed by about 40%. Methods for forcing closure of eddy covariance measurements will be discussed as well as how this might affect the consistency of fluxes