Were it not for the presence of a solar wind, the intrinsic magnetic field of the Earth—if observed from far enough out in space—would appear to be almost perfectly dipolar, with the axis of the dipole presently tilted by some 10° with respect to the rotation axis. At the Earth's surface, the axial dipolar term is dominant, which serves among other uses as a basis for both orientation with a compass and plate-tectonic applications of paleomagnetism. One can ask whether this dipolar dominance would hold were the field observed from just above its source at the core-mantle boundary. The answer is that indeed it does (Figure 1), although of course the dipolar part is reduced in amplitude relative to the higher-order (shorter wavelength) terms. These higher-order terms are compatible with a flat, white-noise-like, power spectrum. This implies that there are similar amounts of energy stored in the various harmonic terms (see, for example,Constable and Parker [1988]), although a “pink” spectrum (that is, slightly more power at the longer wavelengths) is as plausible [Hulot et al., 1992]. Terms beyond degree and order 13 (that is, wavelengths shorter than 1500 km) are contaminated by crustal and lithospheric (that is, surficial static) fields.