Chris: This map is one reconstruction of the world’s major mid-ocean ridges about 83 million years ago. This is recent enough that you can see earlier versions of many of the ridges in the modern ocean basins. The north-south trending ridge on the left-hand side is the mid-Atlantic ridge, ending at a triple junction just south of Greenland; the north-south ridge on the right hand side is the much longer ancestor of the modern East-Pacific rise, the northern half of which has since been subducted beneath North America. These past configurations, and how fast they were producing new oceanic crust, can be worked out relatively easily from the still preserved magnetic anomalies on the modern sea-floor. However, for about 40% of the reconstructed system (marked in grey) things are much more uncertain: not only has the ridge itself been subducted, but the magnetic anomaly record has been too. This raises the question: when calculating global tectonic behaviour, should you use a well-constrained but incomplete record, or fill in the gaps with a possibly wrong best-guess? Chris is trying to better quantify the uncertainties in both methods.