Abstract About 10,000 km of trenches in chasmata and coronae have been identified as possible sites of retrograde subduction on Venus. All the sites have narrow deep trenches elongate along strike with arcuate planforms, ridge-trench-outer rise topographic profiles typical of terrestrial subduction zones, large outer rise curvatures >10 -7 m -1, fractures parallel to the strike of the trench on the outer trench wall and outer rise, and no cross-strike fractures across the trench. Both the northern and southern margins of Latona Corona are possible subduction sites. Identification of a major graben between the two principal outer ridges in southern Latona Corona is evidence of back-arc extension in the corona; the amount of extension is estimated to be more than 2-11 km. The moment exerted by the ridges of southern Latona Corona is insufficient to bend the lithosphere into the observed outer rise shape; a negatively buoyant subducted or underthrust slab is needed. Depending on the unknown trench migration rate, lithospheric subduction can make a significant contribution to mantle cooling on Venus. Venusian chasmata could have a dual character. They may be propagating rifts near major volcanic rises, and subduction trenches far from the rises in the lowlands. Subduction and rifting may occur in close proximity on Venus, unlike on Earth. Rifting induced by hotspots on Venus may be necessary to break the lithosphere and allow subduction to occur. Such a process could result in gradual lithospheric subduction or global, episodic overturn of the lithosphere.