In contrast to pillow lavas from the marine environment around Iceland and Hawaii, tholeiitic pillow lavas from the Mid-Atlantic Ridge at 45° N do not show a correlation between vesicle size and depth of recovery. Post-eruptive vertical block faulting may have displaced the lavas from their original extrusion depth by as much as 1500 m, explaining in part the lack of correlation. Calculation of the original depth of extrusion by compensation for these vertical movements indicates that many of the basalts now high on the Crest Mountains were extruded in deep water (2900–3500 m), possibly within the floor of the Median Valley. Nevertheless, they were able to produce numerous, comparatively large vesicles, permitting the basalts to outgas radiogenic argon trapped in the magmas.Bulk chemical differences between the 45? N basalts and those of Iceland and Hawaii cannot explain why those from 45? N are able to outgas under high hydrostatic pressure. Different histories prior to extrusion may contribute to these varying characteristics.