The magnetic properties of oceanic basalt samples from 55 DSDP sites (301 samples) and 112 dredge haul sites (311 samples) show wide variations, indicative of heterogeneity in the basaltic layer. Remanent intensity and Königsberger ratio are log-normally distributed and range over three orders of magnitude for each collection. For DSDP samples between-site variation is greater than within-site variation, suggesting horizontal variation of magnetization within the source layer of magnetic anomalies. Crustal magnetization estimates from inversion of short-wavelength oceanic magnetic anomalies may therefore be in error. By using arithmetic means rather than the geometric means appropriate for log-normal distributions, estimates of the magnetization of Layer 2 of around 4 × 10−3 Gauss are obtained for the Indian, Pacific and north Atlantic oceans and Caribbean Sea, while lower averages are found in the south Atlantic and Antarctic oceans. Using this as a representative value for abyssal areas, the thickness of the magnetic source layer is greater than commonly assumed, and probably more than a kilometre. This interpretation is supported by DSDP data from some individual sites, notably the deep holes of Leg 37.
The DSDP median destructive fields average only 123 Oe. Although many sites have much higher stabilities than this, there is evidence for unstable NRM at many sites where samples exhibit a strong tendency to acquire viscous remanence components. The low coercivity secondary magnetizations confuse identification of the stable direction during magnetic cleaning. Acquisition of high coercivity secondary remanence (CRM) as a result of maghemitization, which is extensive in the DSDP collection, may also have altered the original NRM direction.
Remanent inclination data from four Tertiary DSDP sites (ages 38–40 Ma) have been combined to give a tentative Eocene VGP for east Antarctica at 68°S 64°E. It has not been possible to compare DSDP data in this way at other sites. Palaeolatitudes deduced from DSDP basalt remanent inclinations show large discrepancies from values expected from plate reconstructions at many sites. Incompletely removed secondary magnetizations, tectonic effects and palaeo-reconstruction errors, alone or in combinations, may be responsible for these discrepancies. A major source of deviation from the expected direction may be the rapid cooling of the basalt which does not allow adequate averaging of secular variation.