We have examined three very long magnetic field profiles taken over ocean basins for the presence of intermediate wavelength magnetic anomalies. One profile was from the Atlantic Ocean in the Transatlantic Geotraverse area, one ran along latitude 35¿S in the SE Pacific, and one ran along 150¿W in the Pacific. All three profiles show the presence of intermediate wavelength (65--1500 km) magnetic anomalies generated in the crust or upper mantle. The analysis of magnetic field power spectra shows that the core field becomes unimportant at about a wavelength of 1500 km. Sea floor spreading anomalies should produce a maximum in power at about a wavelength of 65 km. Between these two wavelengths there should be a minimum in power which is not seen on observed records. Inverting the anomalous field to obtain some idea of the magnetization necessary to explain these intermediate wavelength magnetic anomalies shows that values of magnetization in excess of 1 A m-1 are needed if the magnetized layer is as thick as the ocean crust. Alternatively, rather large thicknesses of upper mantle material with lower intensities of magnetization need to be used. The reason why such magnetization variations exist is not known. It can be shown that upward continuation of the magnetic anomaly signature to an altitude of 350 km (about the perihelion altitude of MAGSAT) will produce anomalies up to 10 nT in amplitude. These should be capable of being seen by MAGSAT, and thus allow us to determine the spatial arrangement of the intermediate wavelength anomalies and hence, hopefully, a clue as to their origin.