Abstract During the SUBPSO1 cruise, seven submersible dives were conducted between water depths of 5350 and 900 m over the collision zone between the New Hebrides island arc and the d'Entrecasteaux Zone (DEZ). The DEZ, a topographic high on the Australian plate, encompasses the North d'Entrecasteaux Ridge (NDR) and the Bougainville guyot, both of which collide with the island-are slope. In this report we use diving observations and samples, as well as dredging results, to analyse the geology of the Bougainville guyot and the outer arc slope in the DEZ-arc collision zone, and to decipher the mechanisms of scamount subduction. These data indicate that the Bougainville guyot is a middle Eocene island arc volcano capped with reef limestones that appear to have been deposited during the Late Oligocene to Early Miocene and in Miocene-Pliocene times. This guyot possibly emerged during the Middle and Late Miocene, and started to sink in the New Hebrides trench after the Pliocene. The rocks of the New Hebrides arc slope, in the collision zone, consist primarily of Pliocene-Recent volcaniclastic rocks derived from the arc, and underlying fractured island-arc volcanic basement, possibly of Late Miocene age. However, highly sheared, Upper Oligocene to Lower Miocene nannofossil ooze and chalk are exposed at the toe of the arc slope against the northern flank of the NDR. Based on a comparison with cores collected at DSDP Site 286, the ooze and chalk can be interpreted as sediments accreted from the downgoing plate. East of the Bougainville guyot an antiform that developed in the arc slope as a consequence of the collision reveals a 500-m-thick wedge of strongly tectonized rocks, possibly accreted from the guyot or an already subducted seamount. The wedge that is overlain by less deformed volcaniclastic island-arc rocks and sediments includes imbricated layers of Late Oligocene to Early Miocene reef and micritic limestones. This wedge, which develops against the leading flank of the guyot, tends to smooth its high-drag shape. A comparison between the 500-m-thick wedge of limestones that outcrops southeast of the guyot and the absence of such a wedge over the flat top of the guyot, although the top is overthrust by island-arc rocks and sediments, can be interpreted to suggest that the wedge moves in the subduction zone with the guyot and facilitates its subduction by streamlining.