Over the past 100??yr, several hypotheses have been proposed for the origin and age of the Great Sand Dunes. These hypotheses differ widely in the descriptions of dune morphometry, the immediate source of eolian sand, and when sand transport occurred. The primary purpose of this paper is to evaluate these hypotheses and, where warranted, to present new ideas about the origin and age of the Great Sand Dunes. To evaluate the previous hypotheses, we had to develop more detailed information about the surficial geology of the northern San Luis Valley. Thus, we mapped the surficial geology of an area extending several tens of kilometers north, south, and west of the Great Sand Dunes and examined subsurface stratigraphy in more than 200 wells and borings. In addition, we used relative-dating criteria and several radiocarbon and OSL ages to establish the chronology of surficial deposits, and we determined the U-Pb ages of detrital zircons to obtain information about the sources of the sand in the Great Sand Dunes. The first principal finding of this study is that the lower part of the closed basin north of the Rio Grande, referred to here as the sump, is the immediate source of the sand in the Great Sand Dunes, rather than the late Pleistocene flood plain of the Rio Grande (the most widely accepted hypothesis). A second principal finding is that the Great Sand Dunes are older than late Pleistocene. They postdate the draining of Lake Alamosa, which began ??? 440??ka, and predate the time when streams draining the west flank of the Sangre de Cristo Mountains were deflected by incipient dunes that formed near the mountain front. Geomorphic and stratigraphic evidence indicate that this deflection occurred prior to the end of the next to last glaciation (Bull Lake), i.e., prior to ??? 130??ka.