The depth to the axial magma chamber at three oceanic spreading centers, as indicated by seismic reflection surveys, is greater than that predicted by conductive cooling thermal models of ridge crests. The additional cooling at the ridge axis is owed to the circulation of seawater through the shallow crust. A theoretical model for the temperature distribution at spreading centers, which includes distributed heat sources and sinks, is presented. By representing the hydrothermal heat loss as a series of heat sinks, the depths to the axial magma chamber at the East Pacific Rise at latitude 9ΓΈN., the southern Juan de Fuca Ridge, and the Lau Basin spreading center are modeled. Heat sinks are added to the input model to increase the solidus depth until the computed magma chamber depth matches the magma chamber depth determined by seismic reflection surveys. The thermal modeling indicates that the axial hydrothermal heat flux is between 7.1 x 10 6 and 13.8 x 10 6 MJ/m 2, or 10 to 20% of the total missing heat at the ridge axis, requiring extensive low-temperature circulation off axis. High-temperature vents that are spaced 1 km apart and are active between 4 and 10% of the time would account for the axial hydrothermal heat loss. surface heat flow and the elevation of the crust can also be estimated [Davis and Lister, 1974; Oldenberg, 1975; Trβ’hu, 1975; Lister, 1977]. Theoretical models predict that both heat flow and elevation vary linearly with the square root of the age of the crust. Away from mid-ocean ridges, excellent agreement is found between the theoretical predictions and the actual heat flow and depth versus age curves [Trβ’hu, 1975; Parsons and Sclater, 1977; Sclater et al., 1980]. Measured heat flow near spreading centers is much lower than predicted by purely conductive models, suggesting that near the ridge axis additional heat is removed by circulation of sea water through the hot, newly formed crust [Lister, 1972; Anderson and Hobart, 1976; Sleep and Wolery, 1978]. The discovery of hot springs at the Galapagos Ridge at longitude 86ΓΈW. [Corliss et al., 1979] and the East Pacific Rise (EPR) at latitude 21ΓΈN. [RISE Project Group, 1980] confirmed the hypothesis that hydrothermal circulation is occurring at ridges. Estimates for the heat loss of high temperature (350 ΓΈ C) vents on the EPR range from 250 MW for a single vent [Macdonald et al., 1980] to 170 to 210 MW for a cluster of vents [Converse et al., 1984]. These results emphasize the need to include heat loss due to hydrothermal circulation in theoretical thermal models if accurate temperature distributions near the ridge axis are desired.