The discovery that methane from dissociation of gas-hydrates could be an important factor in the global carbon cycle resolves the major discrepancy in estimates of the increase of terrestrial biomass from the Last Glacial period to the present. Carbon isotope budgeting using the marine carbon isotopic record results in an estimate centered around 500 GtC , whereas palaeovegetation reconstruction (with biosphere models) gives averages around 1000 GtC . The discrepancy may be resolved by considering release of isotopically light methane through destabilization of gas hydrates. This provides a unique means of estimating the contribution of gas hydrates to the deglacial rise in atmospheric methane. A release of ∼120 GtC methane, makes a biospheric carbon transfer of ∼1000 GtC compatible with the marine carbon isotope data. This, however, represents less than 30% of the enhanced atmospheric methane production between 18 and 8 ka observed in ice cores, supporting the theory that glacial–interglacial variations in atmospheric methane were driven primarily by changes in the extent of tropical and temperate wetlands and not by methane release from clathrates. By balancing the deglacial carbon budget we demonstrate that global carbon models will have to incorporate glacial–interglacial vegetation shifts of at least 1000 GtC , which many currently find difficult.