IT is now generally accepted that pterosaurs, Mesozoic reptiles, were true fliers, but the nature of their flight apparatus is still much disputed. Evidence has been presented in favour of bird-like reconstructions with narrow, stiff wings free of the legs(1-6) and bat-like reconstructions with extensive wings incorporating both fore and hind limbs(7-10), but the Solnhofen Limestone pterosaurs, upon which these models are based, are not sufficiently well preserved to resolve these conflicting interpretations. Here we present a new model, founded on Sordes pilosus from the Jurassic of middle Asia (ref. 11, and N.N.B. and D.M.U., manuscript submitted), in which exceptionally well preserved wing membranes show that the hind limbs of pterosaurs were intimately involved in the flight apparatus; connected externally to the main wing membrane and internally by a uropatagium, controlled by the fifth toe. Sordes also reveals that, uniquely among flying vertebrates, pterosaurs had a structurally non-homogenous flight surface with a stiffened outer half and a softer, more extensible inner region.