We estimate temperature anomalies and water content in the mantle transition zone from the depth of the "660-km discontinuity" and tomographically determined P-velocity anomalies. We assume a linear dependence of the discontinuity depths and P-velocity on temperature anomaly and water content. Beneath the Philippine Sea, where the Pacific plate is subducted, temperature anomalies are as low as -500K to -700 K within and near the stagnant Pacific slab and the water content is estimated to be in the range of 1-1.4 wt.%H2O. The west Philippine basin, away from the Pacific slab, does not have a significant temperature anomaly or water content. Beneath western Japan, where the Pacific slab is subducted, we obtain temperature anomalies up to -300 to -600 K and water content up to 1-1.5 wt.%H2O. Many problems remain to be solved for obtaining a definitive conclusion on the presence of water and quantitative estimates of temperature and water content. Estimates of the temperature anomaly and water content are highly sensitive to the input seismic parameters (the discontinuity depths and P-velocities) and the assumed dependence of the seismic parameters on temperature and water content determined from experimental studies. More accurate estimates of the seismic parameters and the experimental data measured under the pressure and temperature conditions of the mantle transition zone are necessary. Anelastic attenuation is probably enhanced by water, which might break down the linear dependence of the discontinuity depths and P-velocity on temperature anomaly and water content. A non-linear optimization approach using better seismic and experimental data may be required for obtaining more conclusive evidence for the presence of water.