The western Canaries, relatively little studied until a few years ago from the geological point of view, have however provided decisive data for understanding many of the most important geological problems of the Archipelago, which would probably have been dilucidated earlier, had the study begun with the most recent islands, as occurs in similar chains of oceanic volcanic islands in other parts of the world. To summarize the main geological features and evolutionary characteristics of both islands we emphasize the following stages of development: During the Pliocene, a submarine volcanic edifice or seamount formed in the island of La Palma, made up of pillow lavas, pillow breccias and hyaloclastites, intruded by trachytic domes, plugs of gabbros, and a highly dense dyke swarm. The intense magmatic and dyke intrusion uplifted the seamount up to 1500 m, tilting it 45-50º to the SW. This intrusive phase was followed by a period of quiescence and erosion of the emerged submarine edifice. The definitive consolidation and progression of the construction of the island continued from at least 1.77 ma in angular and erosive discordance over the submarine basement. The subaerial volcanic reactivation, in which explosive volcanism predominated during the initial stages, producing abundant volcanoclastic and phreatomagmatic materials at the base of the subaerial edifice, persisted in a highly continuous manner until at least 0.41 ma. This initial subaerial stage shaped the northern volcanic shield, formed by the accumulation of several superimposed volcanoes, approximately concentric in relation to one another and the submarine basement. The initial stage of the northern volcanic shield lasted between 1.77 and 1.20 ma, during which period the Garafía volcano was built to a height of 2500-3000 m, with steeply sloping flanks, formed predominantly by alkaline basalts with abundant pahoehoe lavas. The rapid growth and progressive instability of the Garafía volcano culminated some 1.20 ma ago in a gravitational landslide of the south flank of the volcanic edifice. The eruptive activity that followed the collapse built the Taburiente volcano, that rests upon a clear angular discordance caused by the landslide. The landslide depression was filled completely some 0.89 ma ago, as shown by the age of the first lavas to overflow the collapse embayment. The filling-in of the depression by the Taburiente volcano lavas finally formed a sequence of horizontal lavas, predominantly alkaline basalts, that ponded against the headwall of the landslide scarp forming a plateau in the centre of the volcanic shield. Coinciding approximately with the Matuyama-Brunhes boundary (0.78 ma) an important reorganisation of the Taburiente volcano took place, the dispersed emission centres of which progressively concentrated in three increasingly defined rifts (NW, NE and N-S) and subsequently in a central edifice situated at the geometrical centre of the volcanic shield. The abundant emissions of this final stage covered the earlier formations with sequences of lava flows up to 1000 m thick in places, with the exception of a part of the alignments of cones of the rifts. The basaltic lavas evolved towards more differentiated phonolitic and trachytic terms at the terminal phases of construction of the volcano. One of these rifts, the southern or Cumbre Nueva rift, developed more than the others, possibly because the volcanism already began to migrate southwards, forming a N-S trending dorsal ridge over 2500 m high. The progressive instability of the Cumbre Nueva rift, due to overgrowth, triggered the gravitational landslide of the western flank, in a process that took place about 560 ka ago, involving the detachment of some 180-200 km3 and the formation of a wide depression (the Valle de Aridane) and the beginning of the formation, by incision and retrogressive erosion, of the Caldera de Taburiente. The activity subsequent to the collapse in the northern shield was preferentially concentrated in the interior of the new collapse basin, quickly building the Bejenado stratovolcano. This activity was coetaneous with that of other residual centres dispersed over the flanks of the shield. The initially basanite lavas of Bejenado volcano evolved to mafic tephrites in differentiated lateral and terminal vents. The activity of the volcanic shield ceased definitively some 0.4 ma ago. After a transition period with a certain degree of activity associated with Bejenado late peripheral vents, volcanism was definitively located until the present in the new Cumbre Vieja volcano, at the south of the island. The oldest Cumbre Vieja lavas have been dated in 123 ka, although the first eruptions of the volcano may be considerably older. During this last stage of volcanism in La Palma a N-S trending rift has been formed, with predominantly basanitic, tephritic and tephri-phonolitic lavas, and intrusions of domes of tephri-phonolites and phonolites, frequently associated with eruptive vents. Numerous submarine eruptive vents, several of which are apparently very recent, have recently been observed and sampled at the prolongation of the Cumbre Vieja rift southwards in the ocean. The foreseeable geological evolution of this rift is similar to that of its Cumbre Nueva predecessor, towards a progressive development and increasing instability, although changes may take place that may modify it towards more stable configurations, fundamentally the submarine progression of the southern tip of the rift, that could redistribute the volume of emitted materials, reduce the aspect ratio of the volcano and, consequently, its instability. The en echelon faults generated during the 1949 eruption have been interpreted as a possible detachment of the western flank of the volcano, although a more favourable hypothesis would be that such faults are surficial and contribute to accommodating the volcano by reducing its instability. A noteworthy aspect is the important role played by the mobility of the general feeding system of the volcanism in shaping the form and structure of the island. If the volcanism had not continually migrated southward since the final stages of construction of the northern shield, the island of La Palma would probably have taken on a similar configuration to that of the islands of El Hierro or Tenerife, in the shape of a triangular pyramid, with triple-armed rifts and landslide lobes between the rifts. The southward migration of volcanism in La Palma left the northern shield extinct, the rifts incomplete and finally configured an island lengthened in a N-S direction. Another point of interest is that the islands of La Palma and El Hierro are the first of the Canaries to form simultaneously, with possibly alternating eruptive activity, at least in the most recent period. This separation in a "dual line" of islands and the greater depth of its oceanic basement account for the long time they have required to emerge since the formation of the prior island of La Gomera. The island of El Hierro is geologically somewhat younger than La Palma and, because it formed over a stationary source of magma, it presents, in comparison, a perfect, concentric development, with superimposed volcanoes and a regular three-armed rift geometry. The activity of the subaerial volcanism began in El Hierro with the development of Tiñor volcano on the NE flank of the island (approximately from 1.12 to 0.88 ma), with the emission of massive typical basalts. The volcano developed quickly, with different stages of growth, the eruption of Ventejís volcano being the terminal explosive stage, and probably the precursor of the collapse of the NW flank of the edifice some 882 ka ago. The emissions of the new volcano -El Golfo, approximately 545 to 176000 ka- totally filled the depression of the lateral collapse of Tiñor volcano, the lava flows of which then spilled over the flanks of the earlier volcano. The beginning of the construction of the El Golfo volcano seems to have taken place after a relatively long period of activity, probably coinciding with the maximum development of the Cumbre Nueva rift on La Palma. The initial subaerial activity at El Golfo was characterised by basaltic lavas that evolved to trachybasalts and trachytes, and finally towards more differentiated eruptive episodes indicative of the terminal state of the volcanic activity of the El Golfo volcano. The excessive growth of this volcano triggered the failure of its north flank, generating the spectacular scarp and present El Golfo depression. Subsequent volcanism, from emission vents arranged in a three-armed rift system (rift volcanism, with ages ranging from 145 ka to 2500 years, with probably prehistoric eruptions), implies the much more moderate continuation of the earlier predominantly basanitic-tephritic volcanic activity. This period may correspond to that of maximum development of the Cumbre Vieja rift, in the island of La Palma.