Interpretation of the noble gas (Ne, Ar, Kr and Xe) isotopic composition of Ocean Island and Mid Ocean Ridge basalt (OIB and MORB) glass is complicated by the near surface addition of atmosphere-derived noble gases. Although the noble gas elemental composition in these samples is often similar to modern air, equilibration between a seawater vapor phase and a basalt melt can produce similar values in the melt phase [Patterson et al., Geophys. Res. Lett. 17 (1990) 705–708]. With the assumption that laboratory handling must reverse any surficial modern air contamination, it has often been concluded that the measured ‘air-like’ noble gases are introduced into MORB and OIB glass by variable interaction of the magma with seawater, either before or during eruption. Harrison et al.[Earth Planet. Sci. Lett. 171 (1999) 199–207] have recently demonstrated that both MORB popping rock and subglacially erupted Iceland OIB glass contain a component with 130Xe/ 22Ne indistinguishable from modern air. 84Kr/ 22Ne and 36Ar/ 22Ne are also indistinguishable from modern air for air values of 20Ne/ 22Ne. Assuming that the mode of this ‘air-like’ component addition is the same in both systems we show that calculated fractionation values for melt/seawater and melt/freshwater–glacier systems respectively cannot produce the observed elemental abundance pattern. We argue that the dominant source of ‘air-like’ noble gases in both the MORB and OIB is contamination by unfractionated modern air during sampling or sample preparation within the laboratory. We show that this is related to vesicularity and does not appear to be fully removed by current laboratory techniques. Because vesicularity is related to eruption depth and volatile content [J.G. Moore, Nature 282 (1979) 250–253], recently observed correlations between atmosphere-derived noble gases with radiogenic Pb, water content or other trace element pairs [Sarda et al., Science 283 (1999) 666–668; Bach and Niedermann, Earth Planet. Sci. Lett. 160 (1998) 297–309] may not reflect a relationship between these species in the magmatic source. Similarly, the conclusion from the existing MORB and OIB data set that the terrestrial mantle heavy noble gases have many similar features to modern air [Ozima and Igarashi, Earth Planet. Sci. Lett. 176 (2000) 219–232] is premature. These observations can be equally well accounted for by eruption style related vesicularity and related air contamination.