Paleomagnetic results from abyssal igneous cores collected during the Deep-Sea Drilling Project (DSDP) are compared to reference data sets from lava sequences for the period between 13 and 2 m.y. ago in Iceland. Statistical modeling shows that the paleomagnetic data from Iceland can be modeled approximately by a virtual geomagnetic pole distribution which is 84% Fisherian and 16% random. Transforming this result into values of inclination which can be expected at other latitudes provides a measurement of the probability that negative inclinations can occur during times of normal polarity in the northern hemisphere. It appears that inclination of remanent magnetism from holes within 30¿ of the paleoequator cannot provide reliable measurements of polarity. Comparison of paleomagnetic inclinations in basalt from DSDP leg 37 with our model shows that factors other than the geomagnetic field behavior almost certainly were the cause of the measured inclinations. Postemplacement tectonic rotation around an axis parallel to the ridge crest is therefore inferred. Data from leg 46 contain too few measurements for a reliable comparison to be made.