The upper Middle Jurassic-Lower Cretaceous Gravina belt lies along the eastern margin of the Alexander terrane in southeastern Alaska. This group of turbidites and mafic to intermediate volcanic rocks was deformed during mid to Late Cretaceous time during the closing of a basin of unknown size between the Alexander terrane on the west and the Stikine terrane to the east. Therefore structures of Gravina belt rocks largely reflect the final accretion and subsequent transport of the Alexander terrane. Six geologic transects across the central Gravina belt (southern Mitkof Island to northern Douglas Island) define a structural history that includes (1) syndepositional soft-sediment folding and faulting, possibly in conjunction with block tilting and extension; (2) tight to isoclinal folding or thrust faulting, with formation of a slaty cleavage (S₁) striking ∼330°; (3) local coaxial refolding with formation of crenulation cleavage (S₂); (4) development of domainal crenulation folds and cleavage (S₃) oriented at a large angle to the margin of the belt; (5) intrusion of tonalitic plutons around 90 Ma; and (6) right-lateral strike-slip displacement on faults oriented ∼330°. Finite strain measurements on sedimentary rocks suggest the belt was at least twice its present width, normal to the foliation, before deformation. Subhorizontal margin-parallel fold axes, margin-parallel slaty cleavage, and margin-perpendicular stretching lineations suggest orthogonal contraction of the Gravina basin, assuming that oblique plate convergence was not decoupled along strike-slip faults. After contractional deformation, strike-slip faults indicate dextral displacement (probably of the order of several tens of kilometers) of the Alexander terrane with respect to the terranes to the east. Domainal crenulation folds and cleavages at a high angle to the margin of the belt suggest that the change in convergence directions occurred while the rocks (presently at the surface) could still plastically deform. This kinematic interpretation of structures is consistent with changes in plate motions [Engebretson et al., 1985]. Before 100 Ma, the convergence directions between the Kula and North America plates were at a high angle to the continental margin, whereas after 100 Ma, convergence directions were at a small angle to the continental margin. In addition, after 100 Ma, the Kula-North America, and not the Farallon-North America, convergence direction is most compatible with a N-S principal paleostress orientation derived from inversion of strike-slip fault data. This relationship suggests that it may have been the Kula plate that drove northward transport of the Alexander terrane along the margin of North America.