Methane hydrates – natural gas held in clathrate solids that resemble water ice – that occur in sea-floor sediments are on the one hand a potential energy resource and on the other pose great risks. There are between 10¹⁵ to 10¹⁷ m³ buried beneath the ocean floors and an unknown amount in Arctic soils and lakes.  The temperature that confers stability on these peculiar solids depends on pressure. At pressures lower than those at a water depth of around 250m they are unstable. Clathrate crystals form from natural gas and water in sediments at 0°C at that depth and at progressively higher temperatures at deeper levels beneath the seafloor, until geothermal heat flow at a depth of around 2.5 km results in temperatures above about 20°C when they cannot form; there is a depth-temperature window in which gas hydrates may be found in seafloor sediments, which depends on the temperature of deep water. Little is known about the stability of gas hydrates. In some areas there is a steady release of methane that bubbles to the surface, whereas in others they can be detected by seismic surveys in huge volumes that appear to be stable with no release. One area rich in gas hydrates occurs at the continental edge off the Norwegian coast (the Storegga in Norwegian). Periodically sediments at the Storegga fail in massive sub-sea landslides which have resulted in tsunamis in the North Sea. The last such tsunami occurred around 6100 BCE after a slide displaced 3500 km³ of debris, devastating the east coast of Scotland. Either an earthquake triggered the slide or it was due to destabilizing of the clathrates. Either way huge amounts of methane would have been released. At the end of the Palaeocene Era (55 Ma) a global carbon-isotope anomaly coincides with evidence for very rapid climatic warming, which suggests that vast amounts of methane – a far more powerful greenhouse gas than CO₂ – were released from submarine gas hydrates. In recent years the loss without trace of several large ships may have resulted from a lowering in the density of surface water by gas bubbles that caused the vessels to founder. One country that plans to exploit gas hydrates off its Pacific cast is Japan, and recent surveys indicate a large basin underlain by highly disturbed sediments which contain clathrates on the flank of the basin (Bangs, N.L. et al. 2010. Massive methane release triggered by seafloor erosion offshore southwestern Japan. Geology, v. 38, p. 1019-1022). It appears that bottom currents eroded the seafloor to destabilize the clathrates that then ‘erupted’ ripping through the sediments to release around 1.5 x 10¹¹ m³ of methane. Clearly, drilling into gas hydrate deposits is going to be a risky business; drilling will reduce the pressure so that gas is released and it is not known whether or not this might trigger a form of chain reaction. In the longer term, warming of deep water as a result of climate change could place much larger areas of clathrate-rich seafloor in a knife edge.