The structures of Tibet were generated during the accretion on to the Asian plate, firstly of the Qiangtang Terrane during the Triassic, then the Lhasa Terrane during the Jurassic--Cretaceous and finally the Indian continent during the Palaeogene. The southern Kunlun mountains show intense deformation associated with the accretion of deep water sediments on to an active plate margin. The deformation was essentially by footwall propagation of thrusts, though there was pronounced out-of-sequence thrusting with the deformation of basins above the main thrust zone, and the back steepening and backthrusting of earlier structures. The Jinsha Suture probably represents the southern edge of this zone. The Banggong Suture between the Qiangtang and Lhasa Terranes is characterized by pre-collisional ophiolite obduction for over 100 km to the south across the Lhasa Terrane, plus local intense intracratonic deformation of parts of the Lhasa Terrane. However, for this collision there is now very little evidence for intense deformation along the line of the suture and the Qiangtang Terrane itself remained only weakly deformed throughout. Post--Middle Cretaceous, pre-Tertiary deformation of the Lhasa region produced upright- to north-verging folds which decrease in intensity northwards. They may have been formed at the margin of the Gangdise batholith, or they may have originated from early collisional phases along the line of the Indus--Zangbo Suture. However this deformation is approximately synchronous with the more intense deformation of the Xigatse flysch on the accretionary prism and is therefore probably subduction-related, predating collision. Tertiary deformation is relatively widespread across Tibet, producing SSE-directed thrusts across the Fenghuo Shan region of the Qiangtang Terrane and across the northern part of the Lhasa Terrane. Several hundred kilometres shortening can be estimated to have occurred during this deformation, probably reworking older Mesozoic structures. However this shortening is insufficient to provide all of that estimated from palaeomagnetic work or from a study of displacement rates of the Indian plate, and much of the displacement of India into Asia during the Tertiary must be taken up on strike-slip faults in Tibet or on thrusts and strike-slip faults in central Asia north of the Tibetan Plateau. The Tertiary shortening cannot account for all the thickening of the Tibetan crust.