The original hypothesis of Neoproterozoic global glacial conditions, proposed by Joe Kirschvink (California Institute of Technology) and Paul Hoffman (emeritus at Harvard) in the 1990s was that conditions became so severe that the Earth was encased in glacial- and sea ice from pole to pole. As EPN has charted since 2000, that ‘hard’ Snowball variant has become increasingly less favoured by most geoscientists (Kerr, R.A. 2010. Snowball Earth has melted back to a profound wintry mix. Science, v. 327, p. 1186). However, evidence supporting low latitude glaciations continues to emerge (, F.A. and 9 others 2010. Calibrating the Cryogenian. . Science, v. 327, p. 1241-1243). In the latest, diamictites of the so-called ‘Sturtian’ glaciation in north-western Canada are interbedded with volcanic rocks that give a very precise age of 716.5 Ma. That age happens to coincide with outpouring of the regionally massive Franklin flood basalts whose palaeomagnetism gives equatorial latitudes, the first recorded for the Sturtian glaciation: the later Marinoan glaciation (~635 Ma) provides most low-latitude evidence for Snowball conditions. The paper by Francis Macdonald and co-workers also gives detailed carbon isotope data for a continuous sedimentary record from >811 to 583 Ma.

A potential spanner in the works for the entire Snowball Earth hypothesis is the discovery of a strange anomaly concerning palaeomagnetic pole positions during latest Neoproterozoic times  (Abrajevitch, A. & van der Voo, R. 2010. Incompatible Ediacaran paleomagnetic directions suggest an equatorial geomagnetic dipole hypothesis. Earth and Planetary Science Letters, v. 293, p. 164-170). Paleomagnetism from glaciogenic rocks is the lynchpin for the notion of Snowball Earth, some occurrences recording tropical latitudes. Alexandra Abrajevitch (Kochi University, Japan) and Rob van der Voo (University of Michigan) report palaeomagnetic results for igneous rocks between 600 and 550 Ma in what are now North America and Scandinavia. The data show original inclinations of the magnetic field that are both steep and shallow, indicating high and low latitudes respectively. Plotting inclination against radiometric age for what were separate continental masses in the Ediacaran Period reveals repeated rapid changes from high to low palaeolatitudes that simply cannot be accounted for by continental drift: plate tectonic rates would have to have been unaccountably fast (~45 cm yr^-1). To account for the abrupt shifts the authors turn not to true polar wander – due to changes in the geometry of the geomagnetic dipole – but to rapid flips in the orientation of the dipole between a coaxial and an equatorial alignment, perhaps due to dramatic changes of circulation within the liquid outer core. Familiar geomagnetic reversals normally shift the magnetic poles between roughly the geographic pole positions. Yet there are data showing that for brief periods the reversing poles do pass through equatorial latitudes but at very low magnetic field strength. In the cases from the Ediacaran the geomagnetic poles dwelt at tropical latitudes for long periods and maintained a strong field. Were such strange behaviour demonstrated earlier in the Neoproterozoic, during the Cryogenian period of supposed Snowball events, that would undermine the whole basis for the hypothesis. It seems inevitable that geophysicists will scurry to check the earlier palaeomagnetic data, analysing more igneous rocks on all continents at the narrowest possible time intervals.