The Green River Formation was deposited between 53.5 and 48.5 Ma. The Angelo, Fossil Butte, and Lower members of the Green River Formation at Fossil Basin, preserve ash fall tuffs deposited in ancient Fossil Lake. 40Ar/39Ar dating of sanidine yielded eruptive ages of 51.29 ± 1.29 Ma and 52.20 ± 3.08 Ma for two of the tuff beds within Fossil Basin. Immobile element and mineral compositions of Fossil Basin tuffs indicate that most tuffs erupted from a subduction zone originally as rhyolites and dacites. X-ray diffraction analyses reveal that the tuffs' glassy matrices have been altered to illite, calcite, clinoptilolite, analcime, albite, and K-feldspar. The variable alteration of the tuff beds confirms previous studies of Fossil Lake's salinity fluctuation through time. One outcrop (FB-10), which was previously interpreted to represent the K-spar tuff, has biotite of different compositions from that in known K-spar tuff samples (FB-09 and FB-11). Tuff horizons from the Greater Green River Basin have feldspar and biotite compositions similar to those from tuffs in Fossil Basin and are interpreted to have the same eruptive sources. Based on age and proximity, the Absaroka and Challis volcanic fields are the likely sources of tephra deposits in Fossil Basin and the Greater Green River Basin. Calc-alkaline tephras in these lacustrine basins have similar magmatic characteristics to the tuff of Ellis Creek (48.4 ± 1.6 Ma) from the Challis volcanic field. However, major and trace element, and mineral compositions of Absaroka and Challis volcanic rocks are not distinctive enough to definitively determine the source of most Fossil Basin and Greater Green River Basin tephras. Two samples, FB-10 from Fossil Basin and WN-79.15 from the Greater Green River Basin, have compositions similar to calc-alkaline magmas, but have some mineral compositions with A-type chemical affinities; consequently we conclude that they were erupted from volcanoes within the Challis volcanic field. Compositions of Challis volcanic rocks may have important implications for the development of a slab window in western North America during the Eocene. Compositional variation of Challis volcanic rocks through time indicates that calc-alkaline rocks with a slight A-type component erupted early in its history, and as the slab window matured the Challis volcanic field dominantly erupted rocks with a more A-type chemical affinity. A slab window may have developed due to the Farallon slab subducting at a shallow angle beneath the North American plate, and gravity may have caused it to break to the north. Through time the slab could have torn to the south and by 50 Ma the slab window would have been opening beneath the Challis volcanic field. This would have erupted calc-alkaline magmas, but upwelling of the asthenosphere into the mantle wedge (beneath the North American plate) would have introduced A-type magmatism into the magmatic system. By 45 Ma, the slab would have matured and opened sufficiently beneath the Challis volcanic field to replace calc-alkaline magmatism with, first "transitional" magmatism, and then A-type magmatism as evident in the youngest Challis tuffs.