A complex history of ocean floor spreading in the northeast Atlantic is revealed by detailed marine grid surveys of the magnetic field between Reykjanes ridge and the British Isles and between southern Greenland and Labrador. The spreading that separated Greenland from Rockall bank commenced about 60 m.y. B.P. (anomaly 24 time) at about 1.7 cm/yr, then declined to a low of about 0.7 cm/yr at about 30 m.y. B.P. A triple-ridge junction existed over this interval, with the mid-Labrador Sea ridge (Ran ridge) spreading at about 0.6 cm/yr until extinction during the middle Tertiary. Greenland then began to move as a part of the Americas plate, and the resultant change of spreading direction between Greenland and Europe apparently induced the creation of transform faults spaced 100 km apart. Spreading rates were high again (1.1 cm/yr) by 10 m.y. B.P.; the acceleration phase between 30 and 10 m.y. B.P. is poorly established. Transform faults began to disappear during this phase by gradual rotation of spreading axes, beginning about 40 m.y. B.P. in the north and gradually extending to the south. By 10 m.y. B.P., the Reykjanes ridge had become a continuous oblique ridge. Transform fault extinction progressed from northeast to southeast at about 2 cm/yr. Transform faults of short offset may thus form and disappear repeatedly during the life of an ocean basin; the present ridge crest does not necessarily mirror the original break in detail. The reorientation of spreading axes proceeded at about 1° to 4°/m.y. Before 60 m.y. B.P., an older North Atlantic basin, now deeply buried by sediment, apparently extended into the Labrador basin. South of Rockall bank, the 60 m,y. B.P. rejuvenation of spreading is marked by a distinctive ridge, and the younger basement is systematically elevated over the pre-60 m.y. B.P. basement. Magnetic anomalies and topography reveal reorientation of spreading axes and the creation of new transform faults during the period 60–55 m.y. B.P. to the south of the Charlie Gibbs fracture zone. The total offset of the Charlie Gibbs double fracture zone has not changed significantly since 60 m.y. B.P., but the evolution of a short spreading center between the two fractures is not well known. Magnetic anomalies are truncated by the Charlie Gibbs fracture zone; azimuth changes in fracture trend are suggested at 60 and, less certainly, at 38 m.y. B.P. The peculiar spreading history of the northeast Atlantic appears to have been controlled by variable discharge from a mantle convection plume now under Iceland. Asthenosphere flowing outward from the plume could have provided a large component of viscous stress on the plate bottoms in this region. A discharge peak at about 60 m.y. B.P. split Greenland from Europe and poured out large volumes of basalt onto and into the continental edges and the sea floor south of Rockall bank. Discharge declined subsequently, perhaps in concert with that of other plumes, slowing the plate motion. The Labrador and Norwegian ridges became extinct about 30 or 40 m.y. B.P. A discharge renaissance began afterward, more recent events leaving the Iceland plume about 17 and 7 m.y. B.P. to create the steplike topographic profile of the Reykjanes ridge. The oblique shape of the present Reykjanes ridge may derive from longitudinal stress exerted on the crestal area by southward-flowing asthenosphere; alternatively, the flow weakened the crestal area preferentially, steering the cracking toward a new configuration. Rockall and Seabight troughs are proposed to date from the time of the Bermuda discontinuity in the central Atlantic, variously dated at 135 or 110 m.y. B.P. It is suggested that Hatton bank is pre-Grenville in age, Rockall bank is Grenville, and Rockall trough developed along the Caledonian front. Crustal subsidence versus age in the area of the Reykjanes ridge is compared with the published subsidence history for the more typical (less elevated) North Pacific. Compared to the North Pacific standard, crust in the Reykjanes ridge area is anomalously high by values that depend on crustal age and distance from the Iceland plume. Although crust of all ages is higher than coeval Pacific crust, the anomalies are greatest for crust 0 to 20 m.y. in age (+1.3 to +1.8 km at 57°–61°N and +0.1 to +0.6 km at 47°–51°N) and for crust 60 to 40 m.y. in age (+1.5 to +1.8 km at 57°–61°N and +0.5 to +1.2 km at 47°–51°N). Because these times were also apparently characterized by higher plume discharge, it is proposed that a more buoyant, basalt-rich lithosphere is produced during times of intensive plume convection in the mantle.