The Earth’s mantle holds more carbon than its oceans, atmosphere and con- tinents combined, yet the distribution of carbon within the mantle remains uncertain. Our best constraints on the distribution of carbon within the up- per mantle are derived from the carbon-trace element systematics of ultra- depleted glasses and melt inclusions from mid-ocean ridge basalts. How- ever, carbon-trace element systematics are susceptible to modification by crustal processes, including concurrent degassing and mixing, and melt in- clusion decrepitation. In this study we explore how the influence of these processes varies systematically with both the mantle source and melting pro- cess, thereby modulating both global and local carbon-trace element trends. We supplement the existing melt inclusion data from Iceland with four new datasets, significantly enhancing the spatial and geochemical coverage of melt inclusion datasets from the island. Within the combined Iceland dataset there is significant variation in melt inclusion C/Ba ratio, which is tightly correlated with trace element enrichment. The trends in C/Ba- Ba space displayed by our new data coincide with the same trends in data compiled from global ocean islands and mid-ocean ridges, forming a global array. The overall structure of the global C/Ba-Ba array is not a property of the source, instead it is controlled by CO2 vapour loss pre- and post-melt inclusion entrapment; i.e., the array is a consequence of degassing creating near-constant maximum melt-inclusion carbon contents over many orders of magnitude of Ba concentration. On Iceland, extremely high C/Ba (>100) and C/Nb (>1000) ratios are found in melt inclusions from the most depleted eruptions. The high C/Ba and C/Nb ratios are unlikely to be either analytical artefacts, or to be the product of extreme fractionation of the most incompatible elements during silicate melting. Whilst high C/Ba and C/Nb ratios could be generated by regassing of melt inclusions by CO2 vapour, or by mantle melting occurring in the presence of residual graphite, we suggest the high values most likely derive from an intrinsically high C/Ba and C/Nb mantle component that makes up a small fraction of the Icelandic mantle.