Loess is one of the most extensive surficial geologic deposits in midcontinental North America, particularly in the central Great Plains region of Nebraska. Last-glacial-age loess (Peoria Loess) reaches its greatest known thickness in the world in this area. New stratigraphic, geochronologic, mineralogic, and geochemical data yield information about the age and provenance of Peoria Loess, as well as evaluation of recent climate models. Sixteen new radiocarbon ages and recently acquired optically stimulated luminescence ages indicate that Peoria Loess deposition in Nebraska occurred between ca. 25,000 cal yr B.P. and ca. 13,000 cal yr B.P. After ca. 13,000 cal yr B.P. a period of pedogenesis began, represented by the dark, prominent Brady Soil. At some localities, further loess deposition was minimal. At other localities, sometime after ca. 11,000 cal yr B.P., there were additional episodes of loess deposition (Bignell Loess) intermittently throughout the Holocene. The spatial variability of particle size abundances in Peoria Loess shows a northwest-to-southeast fining in Nebraska, consistent with maps of previous workers that show a northwest-to-southeast thinning of loess. These observations indicate that paleowinds that deposited the loess were from the west or northwest and that the source or sources of Peoria Loess lay to the west or northwest. New mineralogical and geochemical data indicate that the most important sources of loess were likely Tertiary siltstones of the White River and Arikaree Groups, silt facies of Pliocene eolian sediments, and small contributions from Pierre Shale. It is likely that fine-grained silts were transported episodically through the Nebraska Sand Hills from Tertiary and Cretaceous bedrock sources to the north, in agreement with a model presented recently. The identification of Tertiary siltstones and silts as the primary sources of loess is consistent with isotopic data presented in a companion paper. Contributions of glaciogenic silt from the Platte and Missouri Rivers were limited to loess zones close to the valleys of those drainages. An earlier computer-based model of global dust generation during the last glacial period did not identify the Great Plains of North America as a significant source of nonglaciogenic eolian silt. However, a refined version of this model does simulate this region as a significant non-glacial dust source during the last glacial period, in good agreement with the results presented here.