A quasi-global compilation of current meter and temperature records, along with some of the published literature, is used to assess the importance of the meridional eddy heat flux in the ocean circulation, and a comparison is made to a uniform global estimate from altimetric data. Eddy fluxes are found to be important, relative to estimated total heat fluxes, in western boundary current areas of all oceans. Eddy heat flux divergences are potentially important in many other locations, but the mooring coverage is inadequate to be more quantitative. The eddy heat flux in the Southern Ocean is of only modest magnitude compared with that in the western boundary currents, but because of the large circumpolar distance, the integrated flux is very important and probably dominant. All these results are generally consistent with independent estimates made from satellite altimeter variability, but the mooring coverage is not adequate to assess the apparent importance of eddy heat fluxes in the tropics as implied by the altimeter. With the exception of the extreme northeastern North Atlantic, where the local wind dominates the variability, there is no evidence for upgradient fluxes. Climate change models must apparently accurately model the eddy mean-flow interactions near western boundary currents, in the high-latitude North Atlantic, in the tropics, and in the Southern Ocean. One can infer that models describing other physical properties such as carbon or nutrients must also adequately represent the eddy contributions to those fluxes as well. ¿ 1999 American Geophysical Union