Intensity and polarity of the geomagnetic field during Precambrian time

There are only 24 Thellier-type paleointensity estimates for Precambrian rocks. Because orogenesis is episodic, these cluster in a few time intervals: 7 between 820 and 1240 Ma, 7 between 1850 and 2215 Ma, 8 between 2450 and 2765 Ma, and 2 earlier Archean results (3470 Ma). Most late Archean-early Proterozoic results are from dikes. Two results are from large intrusions with multiple phases and slow cooling. Late Precambrian results from the slowly uplifted Grenville Province have an even longer cooling history but partial thermoremanent magnetizations (TRMs) can be dated fairly accurately by 40Ar/39Ar geochronology. The geographic distribution is uneven. Most results are from Canada, Greenland and Baltica, with only one study each from Africa and Australia. The African and Australian results are thermal overprints rather than primary TRMs. Numbers of acceptable results are often small and standard deviations large. Despite these limitations, the virtual dipole moments for almost all studies lie within a range 0.5-1.5 times the published 0.3-300 Ma average, and only two results are conspicuously high (>10(23) Am-2). There is no obvious record of onset and growth of a dynamo field in the Archean or early Proterozoic. A detailed record of a Middle Proterozoic polarity transition has all the expected features: 180degrees reversal along a great circle path, precursor excursion on the same path, a major drop in intensity during the reversal, and reduced intensity before and after. Reversal frequency may have been lower than in recent times, with several chrons of length 20-50 myr around 1460-1400 Ma, 1150-1050 Ma, and 1050-820 Ma, but there are also records of 3 myr and shorter chrons in continuous stratigraphic sections. The dipolar nature of the field is debatable. Roughly contemporaneous results confirm the expected dipole dependence of field strength and paleosecular variation on paleolatitude. On the other hand, there is an apparent excess of directional results with low inclinations and paleolatitudes compared to the dipole prediction but this may be due to bias built into the method.