Abstract. The Woodlark triple junction region, a topographically and structurally complex triangular area of Quaternary age, lies east of Simbo Ridge and southwest of the New Georgia island group, Solomon Islands, at the junction of the Pacific, Australian and Solomon Sea plates. SeaMARC II side-scan imagery and bathymetry in conjunction with seismic reflection profiles, 3.5 kHz records, and petrologic, magnetic and gravity data show that tile active Woodlark spreading centre does not extend into this region.South of the triple junction region, the Woodlark spreading centre reoriented at about 2 Ma into a series of short ESEtrending segments. These segments continued to spread until about 0.5 Ma, when the lithosphere on their northern sides was transferred from the Solomon Sea plate to the Australian plate. Simultaneously the Simbo transform propagated northwards along the western side of the transferred lithosphere, forming a trench-trench-transform triple junction located NNW of Simbo island and a new leaky plate boundary segment that built Simbo Ridge.As the Pacific plate approached, the area east of northern Simbo Ridge was tilted northwards, sheared by dominantly right-lateral faults, elevated, and intruded by arc-related magmas to form Ghizo Ridge. Calc-alkalic magmas sourced beneath the Pacific plate built three large strato-volcanic edifices on the subducting Australian plate: Simbo at the northern end of Simbo Ridge, and Kana Keoki and Coleman seamounts on an extensional fracture adjoining the SE end of Ghizo Ridge.A sediment drape, supplied in part from Simbo and Kana Keoki volcanoes, mantles the east-facing slopes of northern Simbo and Ghizo Ridges and passes distally into sediment ponded in the trench adjoining the Pacific plate. As a consequence of plate convergence, parts of the sediment drape and pond are presently being deformed, and faults are dismembering Kana Keoki and Coleman seamounts.The Woodlark system differs from other modern or Tertiary ridge subduction systems, which show wide variation in character and behaviour. Existing models describing the consequences of ridge subduction are likely to be predictive in only a general way, and deduced rules for the behaviour of oceanic lithosphere in ridge subduction systems may not be generally applicable.