State of the planet· The columns of time
Marinoan Snowball Earth
651 Ma → 635 Ma
The second global glaciation, far shorter than the first, and the only one whose end defines a period.
What the evidence shows
The band is drawn sixteen million years wide, but that is the width of our ignorance about when it began rather than the length of the ice: dating the glacial deposits themselves gives a glaciation of at least four million years, against the Sturtian's fifty-six. What is not in doubt is how it ended, and it ended more sharply than anything else in this part of the record. Within a few metres of rock the ice is gone and a warm carbonate sea is depositing on top of the rubble it left. That end, around 635 million years ago, is the base of the Ediacaran, which makes it the only glaciation on the chart to define a period boundary. The first unambiguous animals appear in the world it leaves behind, close enough in time that the two are hard not to connect and hard to connect rigorously.
Bounds and conventions
Its beginning is loose, somewhere after 654.5 million years; its end is one of the best dated boundaries in the Precambrian, fixed near 635 million years by two independent clocks that agree, a rhenium-osmium age in North America and uranium-lead ages from China, Australia and Namibia. When four continents give the same date for the same layer, the layer is global.