This study proposes a new hypothesis: Debris-covered glaciers served as Pleistocene biological refugia. This is based on detailed studies of vascular plant growth on six debris-mantled glaciers, literally around the world, as well as many casual observations also across the globe. We find that such glaciers are quite common and are distributed globally. Using Carbon Glacier, Mount Rainier, U.S.A., as a type locality and case study, we show aspects of the floristic and structural diversity as well as spatial patterns of plant growth on the glacier surface. Migration strategies, root characteristics, and origin and dispersal strategies for vascular plant species are documented. Also reported are special microclimatic conditions in these areas allowing for this remarkable plant ecology. We find that alpine taxa can grow considerably below their usual altitudinal niche due to the cooler subsurface soil temperatures found on glacial debris with ice underneath, and that may have significantly altered the spatial distribution of such flora during full glacial conditions. This in turn creates previously undocumented areas from which alpine, and perhaps arctic, plant species reestablished in post-glacial time. This hypothesis is complementary to both the nunatak hypothesis and tabula rasa theory and possibly helps solve the ongoing controversy between them. ?? 2007 Regents of the University of Colorado.