New data from an Upper Triassic to Lower Cretaceous deep marine succession-the herein reinstated and restricted Gemuk Group-provide a vital piece of the puzzle for unraveling southwestern Alaska's tectonic history. First defined by Cady et al. in 1955, the Gemuk Group soon became a regional catchall unit that ended up as part of at least four different terranes. In this paper we provide the first new data in nearly half a century from the Gemuk Group in the original type area in Taylor Mountains quadrangle and from contiguous rocks to the north in Sleetmute quadrangle. Discontinuous exposure, hints of complex structure, the reconnaissance level of our mapping, and spotty age constraints together permit definition of only a rough stratigraphy. The restricted Gemuk Group is at least 2250 m thick, and could easily be at least twice as thick. The age range of the restricted Gemuk Group is tightened on the basis of ten radiolarian ages, two new bivalve ages, one conodont age, two U-Pb zircon ages on tuff, and U-Pb ages of 110 detrital zircons from two sandstones. The Triassic part of the restricted Gemuk Group, which consists of intermediate pillow lavas interbedded with siltstone, chert, and rare limestone, produced radiolarians, bivalves, and conodonts of Carnian and Norian ages. The Jurassic part appears to be mostly siltstone and chert, and yielded radiolarians of Hettangian- Sinemurian, Pliensbachian-Toarcian, and Oxfordian ages. Two tuffs near the Jurassic-Cretaceous boundary record nearby arc volcanism: one at 146 Ma is interbedded with red and green siltstone, and a second at ca. 137 Ma is interbedded with graywacke turbidites. Graywacke appears to be the dominant rock type in the LowerCretaceous part of the restricted Gemuk Group. Detrital zircon analyses were performed on two sandstone samples using SHRIMP. One sandstone yielded a dominant age cluster of 133-180 Ma; the oldest grain is only 316 Ma. The second sample is dominated by zircons of 130-154 Ma; the oldest grain is 292 Ma. The youngest zircons are probably not much older than the sandstone itself. Point counts of restricted Gemuk Group sandstones yield average ratios of 24/29/47 for Q/F/L, 15/83/2 for Ls/Lv/Lm, and 41/48/11 for Qm/P/K. In the field, sandstones of the restricted Gemuk Group are not easily distinguished from sandstones of the overlying Upper Cretaceous turbidite-dominated Kuskokwim Group. Petrographically, however, the restricted Gemuk Group has modal K-feldspar, whereas the Kuskokwim Group generally does not (average Qm/P/K of 64/36/0). Some K-feldspar-bearing graywacke that was previously mapped as Kuskokwim Group (Cady et al., 1955) is here reassigned to the restricted Gemuk Group. Major- and trace element geochemistry of shales from the restricted Gemuk Group and the Kuskokwim Group show distinct differences. The chemical index of alteration (CIA) is distinctly higher forshales of the Kuskokwim Group than for those of the restricted Gemuk Group, suggesting more intense weathering during deposition of the Kuskokwim Group. The restricted Gemuk Group represents an estimated 90-100 m.y. of deep-water sedimentation, first accompanied by submarine volcanism and later by nearby explosive arc activity. Two hypotheses are presented for the tectonic setting. One model that needs additional testing is that the restricted Gemuk Group consists of imbricated oceanic plate stratigraphy. Based on available information, our preferred model is that it was deposited in a back-arc, intra-arc, or forearc basin that was subsequently deformed. The terrane affinity of the restricted Gemuk Group is uncertain. The rocks of this area were formerly assigned to the Hagemeister subterrane of the Togiak terrane-a Late Triassic to Early Cretaceous arc-but our data show this to be a poor match. None of the other possibilities (e.g., Nukluk and Tikchik subterranes of the Goodnews terrane) is viable; hence, the terrane subdivision and distribution in southwestern Alaska may need