The Magsat data have yielded a global cartography of long‐wavelength anomalies of lithospheric origin. The magnetizations needed to model these anomalies are much higher than those measured on samples by laboratory experiments. An enhancement of both induced and viscous magnetization at high temperature has often been put forward to account for the “missing magnetization”. Experimental measures of viscous magnetization Mvat high temperature are presented for oceanic deep crustal‐ and mantle‐rocks. At room temperature, Mvhereafter defined in experiments as the magnetization acquired after 400 s in a weak field, is negligible in comparison with the room temperature induced magnetization Mk0. At higher temperatures, between 200°C and 450°C, Mvrepresents about 20% of Mk0and is further enhanced near 500°C, ranging from 25 to 130% of Mk0. An extrapolation of these results to geological times shows that the viscous magnetization acquired since the last reversal of the Earth magnetic field may be of the order of magnitude of the induced magnetization Mk0, above 200°C and possibly predominant near 500°C. However, a magnetic uppermost mantle is still needed to explain the magnetizations of oceanic lithosphere derived fr